[PATCH 0/3] Use timespec64 for select like timeouts

[PATCH 0/3] Use timespec64 for select like timeouts

[Deepa Dinamani]

The series is part of y2038 changes.

This changes a few syscalls that have common functions to use
struct timespec64 instead of struct timespec.

This does not include changes to system call uapi interfaces.
Those will be in a different series.

Thanks to Arnd Bergmann for comments on the patches.

Deepa Dinamani (3):
  time: Add missing implementation for timespec64_add_safe()
  fs: poll/select/recvmmsg: use timespec64 for timeout events
  time: Remove timespec_add_safe()

 fs/eventpoll.c         | 12 ++++-----
 fs/select.c            | 67 ++++++++++++++++++++++++++++----------------------
 include/linux/poll.h   | 11 +++++----
 include/linux/time64.h | 17 ++++++-------
 kernel/time/time.c     | 21 ++++++++++++++++
 net/socket.c           |  8 +++---
 6 files changed, 82 insertions(+), 54 deletions(-)

-- 
1.9.1

Cc: John Stultz <john.stultz@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org

[PATCH 1/3] time: Add missing implementation for timespec64_add_safe()

[Deepa Dinamani]

timespec64_add_safe() has been defined in time64.h for
64 bit systems.
But, 32 bit systems only have an extern function prototype defined.
Provide a definition for the above function.

The function will be necessary as part of y2038 changes.
struct timespec is not y2038 safe. All references to timespec will
be replaced by struct timespec64. The function is meant to be a
replacement for timespec_add_safe().

The implementation is similar to timespec_add_safe().

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/time64.h |  4 +---
 kernel/time/time.c     | 25 +++++++++++++++++++++++++
 2 files changed, 26 insertions(+), 3 deletions(-)

diff --git a/include/linux/time64.h b/include/linux/time64.h
index 367d5af..1778937 100644
--- a/include/linux/time64.h
+++ b/include/linux/time64.h
@@ -136,13 +136,11 @@ extern void set_normalized_timespec64(struct timespec64 *ts, time64_t sec, s64 n
 
 /*
  * timespec64_add_safe assumes both values are positive and checks for
- * overflow. It will return TIME_T_MAX if the returned value would be
- * smaller then either of the arguments.
+ * overflow. It will return TIME64_MAX in case of overflow.
  */
 extern struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
 					 const struct timespec64 rhs);
 
-
 static inline struct timespec64 timespec64_add(struct timespec64 lhs,
 						struct timespec64 rhs)
 {
diff --git a/kernel/time/time.c b/kernel/time/time.c
index be115b0..558f5fe 100644
--- a/kernel/time/time.c
+++ b/kernel/time/time.c
@@ -769,3 +769,28 @@ struct timespec timespec_add_safe(const struct timespec lhs,
 
 	return res;
 }
+
+#if __BITS_PER_LONG != 64
+
+/*
+ * Add two timespec64 values and do a safety check for overflow.
+ * It's assumed that both values are valid (>= 0).
+ * And, each timespec64 is in normalized form.
+ */
+struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
+				const struct timespec64 rhs)
+{
+	struct timespec64 res;
+
+	set_normalized_timespec64(&res, lhs.tv_sec + rhs.tv_sec,
+			lhs.tv_nsec + rhs.tv_nsec);
+
+	if (unlikely(res.tv_sec < lhs.tv_sec || res.tv_sec < rhs.tv_sec)) {
+		res.tv_sec = TIME64_MAX;
+		res.tv_nsec = 0;
+	}
+
+	return res;
+}
+
+#endif
-- 
1.9.1

[PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Deepa Dinamani]

struct timespec is not y2038 safe.
Even though timespec might be sufficient to represent
timeouts, use struct timespec64 here as the plan is to
get rid of all timespec reference in the kernel.

The patch transitions the common functions:
poll_select_set_timeout() and select_estimate_accuracy()
to use timespec64. And, all the syscalls that use these
functions are transitioned in the same patch.

The restart block parameters for poll uses monotonic time.
Use timespec64 here as well to assign timeout value. This
parameter in the restart block need not change because
this only holds the monotonic timestamp at which timeout
should occur. And, unsigned long data type should be big
enough for this timestamp.

The system call interfaces will be handled in a separate
series.

Compat interfaces need not change as timespec64 is an
alias to struct timespec on a 64 bit system.

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
 fs/eventpoll.c       | 12 +++++-----
 fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
 include/linux/poll.h | 11 +++++----
 net/socket.c         |  8 ++++---
 4 files changed, 54 insertions(+), 44 deletions(-)

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 8a74a2a..10db912 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -1583,15 +1583,15 @@ static int ep_send_events(struct eventpoll *ep,
 	return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false);
 }
 
-static inline struct timespec ep_set_mstimeout(long ms)
+static inline struct timespec64 ep_set_mstimeout(long ms)
 {
-	struct timespec now, ts = {
+	struct timespec64 now, ts = {
 		.tv_sec = ms / MSEC_PER_SEC,
 		.tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC),
 	};
 
-	ktime_get_ts(&now);
-	return timespec_add_safe(now, ts);
+	ktime_get_ts64(&now);
+	return timespec64_add_safe(now, ts);
 }
 
 /**
@@ -1621,11 +1621,11 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 	ktime_t expires, *to = NULL;
 
 	if (timeout > 0) {
-		struct timespec end_time = ep_set_mstimeout(timeout);
+		struct timespec64 end_time = ep_set_mstimeout(timeout);
 
 		slack = select_estimate_accuracy(&end_time);
 		to = &expires;
-		*to = timespec_to_ktime(end_time);
+		*to = timespec64_to_ktime(end_time);
 	} else if (timeout == 0) {
 		/*
 		 * Avoid the unnecessary trip to the wait queue loop, if the
diff --git a/fs/select.c b/fs/select.c
index 8692939..8ed9da5 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -47,7 +47,7 @@
 
 #define MAX_SLACK	(100 * NSEC_PER_MSEC)
 
-static long __estimate_accuracy(struct timespec *tv)
+static long __estimate_accuracy(struct timespec64 *tv)
 {
 	long slack;
 	int divfactor = 1000;
@@ -70,10 +70,10 @@ static long __estimate_accuracy(struct timespec *tv)
 	return slack;
 }
 
-u64 select_estimate_accuracy(struct timespec *tv)
+u64 select_estimate_accuracy(struct timespec64 *tv)
 {
 	u64 ret;
-	struct timespec now;
+	struct timespec64 now;
 
 	/*
 	 * Realtime tasks get a slack of 0 for obvious reasons.
@@ -82,8 +82,8 @@ u64 select_estimate_accuracy(struct timespec *tv)
 	if (rt_task(current))
 		return 0;
 
-	ktime_get_ts(&now);
-	now = timespec_sub(*tv, now);
+	ktime_get_ts64(&now);
+	now = timespec64_sub(*tv, now);
 	ret = __estimate_accuracy(&now);
 	if (ret < current->timer_slack_ns)
 		return current->timer_slack_ns;
@@ -260,7 +260,7 @@ EXPORT_SYMBOL(poll_schedule_timeout);
 
 /**
  * poll_select_set_timeout - helper function to setup the timeout value
- * @to:		pointer to timespec variable for the final timeout
+ * @to:		pointer to timespec64 variable for the final timeout
  * @sec:	seconds (from user space)
  * @nsec:	nanoseconds (from user space)
  *
@@ -269,26 +269,28 @@ EXPORT_SYMBOL(poll_schedule_timeout);
  *
  * Returns -EINVAL if sec/nsec are not normalized. Otherwise 0.
  */
-int poll_select_set_timeout(struct timespec *to, long sec, long nsec)
+int poll_select_set_timeout(struct timespec64 *to, time64_t sec, long nsec)
 {
-	struct timespec ts = {.tv_sec = sec, .tv_nsec = nsec};
+	struct timespec64 ts = {.tv_sec = sec, .tv_nsec = nsec};
 
-	if (!timespec_valid(&ts))
+	if (!timespec64_valid(&ts))
 		return -EINVAL;
 
 	/* Optimize for the zero timeout value here */
 	if (!sec && !nsec) {
 		to->tv_sec = to->tv_nsec = 0;
 	} else {
-		ktime_get_ts(to);
-		*to = timespec_add_safe(*to, ts);
+		ktime_get_ts64(to);
+		*to = timespec64_add_safe(*to, ts);
 	}
 	return 0;
 }
 
-static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
+static int poll_select_copy_remaining(struct timespec64 *end_time,
+				      void __user *p,
 				      int timeval, int ret)
 {
+	struct timespec64 rts64;
 	struct timespec rts;
 	struct timeval rtv;
 
@@ -302,16 +304,18 @@ static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
 	if (!end_time->tv_sec && !end_time->tv_nsec)
 		return ret;
 
-	ktime_get_ts(&rts);
-	rts = timespec_sub(*end_time, rts);
-	if (rts.tv_sec < 0)
-		rts.tv_sec = rts.tv_nsec = 0;
+	ktime_get_ts64(&rts64);
+	rts64 = timespec64_sub(*end_time, rts64);
+	if (rts64.tv_sec < 0)
+		rts64.tv_sec = rts64.tv_nsec = 0;
+
+	rts = timespec64_to_timespec(rts64);
 
 	if (timeval) {
 		if (sizeof(rtv) > sizeof(rtv.tv_sec) + sizeof(rtv.tv_usec))
 			memset(&rtv, 0, sizeof(rtv));
-		rtv.tv_sec = rts.tv_sec;
-		rtv.tv_usec = rts.tv_nsec / NSEC_PER_USEC;
+		rtv.tv_sec = rts64.tv_sec;
+		rtv.tv_usec = rts64.tv_nsec / NSEC_PER_USEC;
 
 		if (!copy_to_user(p, &rtv, sizeof(rtv)))
 			return ret;
@@ -396,7 +400,7 @@ static inline void wait_key_set(poll_table *wait, unsigned long in,
 		wait->_key |= POLLOUT_SET;
 }
 
-int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
+int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time)
 {
 	ktime_t expire, *to = NULL;
 	struct poll_wqueues table;
@@ -522,7 +526,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -545,7 +549,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
  * I'm trying ERESTARTNOHAND which restart only when you want to.
  */
 int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time)
+			   fd_set __user *exp, struct timespec64 *end_time)
 {
 	fd_set_bits fds;
 	void *bits;
@@ -622,7 +626,7 @@ out_nofds:
 SYSCALL_DEFINE5(select, int, n, fd_set __user *, inp, fd_set __user *, outp,
 		fd_set __user *, exp, struct timeval __user *, tvp)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	struct timeval tv;
 	int ret;
 
@@ -648,15 +652,17 @@ static long do_pselect(int n, fd_set __user *inp, fd_set __user *outp,
 		       const sigset_t __user *sigmask, size_t sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 ts64, end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
 		if (copy_from_user(&ts, tsp, sizeof(ts)))
 			return -EFAULT;
+		ts64 = timespec_to_timespec64(ts);
 
 		to = &end_time;
-		if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec))
+		if (poll_select_set_timeout(to, ts64.tv_sec, ts64.tv_nsec))
 			return -EINVAL;
 	}
 
@@ -779,7 +785,7 @@ static inline unsigned int do_pollfd(struct pollfd *pollfd, poll_table *pwait,
 }
 
 static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
-		   struct timespec *end_time)
+		   struct timespec64 *end_time)
 {
 	poll_table* pt = &wait->pt;
 	ktime_t expire, *to = NULL;
@@ -854,7 +860,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -868,7 +874,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 			sizeof(struct pollfd))
 
 int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
-		struct timespec *end_time)
+		struct timespec64 *end_time)
 {
 	struct poll_wqueues table;
  	int err = -EFAULT, fdcount, len, size;
@@ -936,7 +942,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 {
 	struct pollfd __user *ufds = restart_block->poll.ufds;
 	int nfds = restart_block->poll.nfds;
-	struct timespec *to = NULL, end_time;
+	struct timespec64 *to = NULL, end_time;
 	int ret;
 
 	if (restart_block->poll.has_timeout) {
@@ -957,7 +963,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds,
 		int, timeout_msecs)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (timeout_msecs >= 0) {
@@ -993,7 +999,8 @@ SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds, unsigned int, nfds,
 		size_t, sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
diff --git a/include/linux/poll.h b/include/linux/poll.h
index 9fb4f40..37b057b 100644
--- a/include/linux/poll.h
+++ b/include/linux/poll.h
@@ -96,7 +96,7 @@ extern void poll_initwait(struct poll_wqueues *pwq);
 extern void poll_freewait(struct poll_wqueues *pwq);
 extern int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
 				 ktime_t *expires, unsigned long slack);
-extern u64 select_estimate_accuracy(struct timespec *tv);
+extern u64 select_estimate_accuracy(struct timespec64 *tv);
 
 
 static inline int poll_schedule(struct poll_wqueues *pwq, int state)
@@ -153,12 +153,13 @@ void zero_fd_set(unsigned long nr, unsigned long *fdset)
 
 #define MAX_INT64_SECONDS (((s64)(~((u64)0)>>1)/HZ)-1)
 
-extern int do_select(int n, fd_set_bits *fds, struct timespec *end_time);
+extern int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time);
 extern int do_sys_poll(struct pollfd __user * ufds, unsigned int nfds,
-		       struct timespec *end_time);
+		       struct timespec64 *end_time);
 extern int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time);
+			   fd_set __user *exp, struct timespec64 *end_time);
 
-extern int poll_select_set_timeout(struct timespec *to, long sec, long nsec);
+extern int poll_select_set_timeout(struct timespec64 *to, time64_t sec,
+				   long nsec);
 
 #endif /* _LINUX_POLL_H */
diff --git a/net/socket.c b/net/socket.c
index 5dbb0bb..bdfe115 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -2171,7 +2171,8 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 	struct mmsghdr __user *entry;
 	struct compat_mmsghdr __user *compat_entry;
 	struct msghdr msg_sys;
-	struct timespec end_time;
+	struct timespec64 end_time;
+	struct timespec64 timeout64;
 
 	if (timeout &&
 	    poll_select_set_timeout(&end_time, timeout->tv_sec,
@@ -2223,8 +2224,9 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 			flags |= MSG_DONTWAIT;
 
 		if (timeout) {
-			ktime_get_ts(timeout);
-			*timeout = timespec_sub(end_time, *timeout);
+			ktime_get_ts64(&timeout64);
+			*timeout = timespec64_to_timespec(
+					timespec64_sub(end_time, timeout64));
 			if (timeout->tv_sec < 0) {
 				timeout->tv_sec = timeout->tv_nsec = 0;
 				break;
-- 
1.9.1

[PATCH 3/3] time: Remove timespec_add_safe()

[Deepa Dinamani]

All references to timespec_add_safe() now use
timespec64_add_safe().

The plan is to replace struct timespec references
with struct timespec64 throughout the kernel as
timespec is not y2038 safe.

Drop timespec_add_safe() and use timespec64_add_safe()
for all architectures.

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/time64.h | 15 +++++++--------
 kernel/time/time.c     |  4 ----
 2 files changed, 7 insertions(+), 12 deletions(-)

diff --git a/include/linux/time64.h b/include/linux/time64.h
index 1778937..7e5d2fa 100644
--- a/include/linux/time64.h
+++ b/include/linux/time64.h
@@ -65,7 +65,6 @@ static inline struct itimerspec64 itimerspec_to_itimerspec64(struct itimerspec *
 # define timespec64_equal		timespec_equal
 # define timespec64_compare		timespec_compare
 # define set_normalized_timespec64	set_normalized_timespec
-# define timespec64_add_safe		timespec_add_safe
 # define timespec64_add			timespec_add
 # define timespec64_sub			timespec_sub
 # define timespec64_valid		timespec_valid
@@ -134,13 +133,6 @@ static inline int timespec64_compare(const struct timespec64 *lhs, const struct
 
 extern void set_normalized_timespec64(struct timespec64 *ts, time64_t sec, s64 nsec);
 
-/*
- * timespec64_add_safe assumes both values are positive and checks for
- * overflow. It will return TIME64_MAX in case of overflow.
- */
-extern struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
-					 const struct timespec64 rhs);
-
 static inline struct timespec64 timespec64_add(struct timespec64 lhs,
 						struct timespec64 rhs)
 {
@@ -222,4 +214,11 @@ static __always_inline void timespec64_add_ns(struct timespec64 *a, u64 ns)
 
 #endif
 
+/*
+ * timespec64_add_safe assumes both values are positive and checks for
+ * overflow. It will return TIME64_MAX in case of overflow.
+ */
+extern struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
+					 const struct timespec64 rhs);
+
 #endif /* _LINUX_TIME64_H */
diff --git a/kernel/time/time.c b/kernel/time/time.c
index 558f5fe..703306a 100644
--- a/kernel/time/time.c
+++ b/kernel/time/time.c
@@ -770,8 +770,6 @@ struct timespec timespec_add_safe(const struct timespec lhs,
 	return res;
 }
 
-#if __BITS_PER_LONG != 64
-
 /*
  * Add two timespec64 values and do a safety check for overflow.
  * It's assumed that both values are valid (>= 0).
@@ -792,5 +790,3 @@ struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
 
 	return res;
 }
-
-#endif
-- 
1.9.1

Re: [PATCH 1/3] time: Add missing implementation for timespec64_add_safe()

[Deepa Dinamani]

Wanted to add a note here:

> + * Add two timespec64 values and do a safety check for overflow.
> + * It's assumed that both values are valid (>= 0).
> + * And, each timespec64 is in normalized form.
> + */
> +struct timespec64 timespec64_add_safe(const struct timespec64 lhs,
> +                               const struct timespec64 rhs)
> +{
> +       struct timespec64 res;
> +
> +       set_normalized_timespec64(&res, lhs.tv_sec + rhs.tv_sec,
> +                       lhs.tv_nsec + rhs.tv_nsec);
> +
> +       if (unlikely(res.tv_sec < lhs.tv_sec || res.tv_sec < rhs.tv_sec)) {

This check can be reduced to only the first condition if we assume the
timespecs passed in to be normalized.
The current patch maintains the way timespec_add_safe() does it for consistency.

-Deepa

Re: [PATCH 1/3] time: Add missing implementation for timespec64_add_safe()

[John Stultz]

On Fri, Apr 29, 2016 at 9:39 AM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> timespec64_add_safe() has been defined in time64.h for
> 64 bit systems.
> But, 32 bit systems only have an extern function prototype defined.
> Provide a definition for the above function.
>
> The function will be necessary as part of y2038 changes.
> struct timespec is not y2038 safe. All references to timespec will
> be replaced by struct timespec64. The function is meant to be a
> replacement for timespec_add_safe().
>
> The implementation is similar to timespec_add_safe().
>
> Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
> Cc: John Stultz <john.stultz@linaro.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> ---
>  include/linux/time64.h |  4 +---
>  kernel/time/time.c     | 25 +++++++++++++++++++++++++
>  2 files changed, 26 insertions(+), 3 deletions(-)

Looks ok at the first glance. I've queued these up for testing,
however I only got #1 and #3 of the set. Are you hoping these two
patches will go through tip/timers/core or are you looking for acks so
they can go via another tree?

thanks
-john

[RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Deepa Dinamani]

struct timespec is not y2038 safe.
Even though timespec might be sufficient to represent
timeouts, use struct timespec64 here as the plan is to
get rid of all timespec reference in the kernel.

The patch transitions the common functions:
poll_select_set_timeout() and select_estimate_accuracy()
to use timespec64. And, all the syscalls that use these
functions are transitioned in the same patch.

The restart block parameters for poll uses monotonic time.
Use timespec64 here as well to assign timeout value. This
parameter in the restart block need not change because
this only holds the monotonic timestamp at which timeout
should occur. And, unsigned long data type should be big
enough for this timestamp.

The system call interfaces will be handled in a separate
series.

Compat interfaces need not change as timespec64 is an
alias to struct timespec on a 64 bit system.

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
Resending to include John and Thomas on this patch as well.
This is to include this patch also in John's tree.
This will let all 3 patches in the series to merged through the same tree.
 
 fs/eventpoll.c       | 12 +++++-----
 fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
 include/linux/poll.h | 11 +++++----
 net/socket.c         |  8 ++++---
 4 files changed, 54 insertions(+), 44 deletions(-)

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 8a74a2a..10db912 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -1583,15 +1583,15 @@ static int ep_send_events(struct eventpoll *ep,
 	return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false);
 }
 
-static inline struct timespec ep_set_mstimeout(long ms)
+static inline struct timespec64 ep_set_mstimeout(long ms)
 {
-	struct timespec now, ts = {
+	struct timespec64 now, ts = {
 		.tv_sec = ms / MSEC_PER_SEC,
 		.tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC),
 	};
 
-	ktime_get_ts(&now);
-	return timespec_add_safe(now, ts);
+	ktime_get_ts64(&now);
+	return timespec64_add_safe(now, ts);
 }
 
 /**
@@ -1621,11 +1621,11 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 	ktime_t expires, *to = NULL;
 
 	if (timeout > 0) {
-		struct timespec end_time = ep_set_mstimeout(timeout);
+		struct timespec64 end_time = ep_set_mstimeout(timeout);
 
 		slack = select_estimate_accuracy(&end_time);
 		to = &expires;
-		*to = timespec_to_ktime(end_time);
+		*to = timespec64_to_ktime(end_time);
 	} else if (timeout == 0) {
 		/*
 		 * Avoid the unnecessary trip to the wait queue loop, if the
diff --git a/fs/select.c b/fs/select.c
index 8692939..8ed9da5 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -47,7 +47,7 @@
 
 #define MAX_SLACK	(100 * NSEC_PER_MSEC)
 
-static long __estimate_accuracy(struct timespec *tv)
+static long __estimate_accuracy(struct timespec64 *tv)
 {
 	long slack;
 	int divfactor = 1000;
@@ -70,10 +70,10 @@ static long __estimate_accuracy(struct timespec *tv)
 	return slack;
 }
 
-u64 select_estimate_accuracy(struct timespec *tv)
+u64 select_estimate_accuracy(struct timespec64 *tv)
 {
 	u64 ret;
-	struct timespec now;
+	struct timespec64 now;
 
 	/*
 	 * Realtime tasks get a slack of 0 for obvious reasons.
@@ -82,8 +82,8 @@ u64 select_estimate_accuracy(struct timespec *tv)
 	if (rt_task(current))
 		return 0;
 
-	ktime_get_ts(&now);
-	now = timespec_sub(*tv, now);
+	ktime_get_ts64(&now);
+	now = timespec64_sub(*tv, now);
 	ret = __estimate_accuracy(&now);
 	if (ret < current->timer_slack_ns)
 		return current->timer_slack_ns;
@@ -260,7 +260,7 @@ EXPORT_SYMBOL(poll_schedule_timeout);
 
 /**
  * poll_select_set_timeout - helper function to setup the timeout value
- * @to:		pointer to timespec variable for the final timeout
+ * @to:		pointer to timespec64 variable for the final timeout
  * @sec:	seconds (from user space)
  * @nsec:	nanoseconds (from user space)
  *
@@ -269,26 +269,28 @@ EXPORT_SYMBOL(poll_schedule_timeout);
  *
  * Returns -EINVAL if sec/nsec are not normalized. Otherwise 0.
  */
-int poll_select_set_timeout(struct timespec *to, long sec, long nsec)
+int poll_select_set_timeout(struct timespec64 *to, time64_t sec, long nsec)
 {
-	struct timespec ts = {.tv_sec = sec, .tv_nsec = nsec};
+	struct timespec64 ts = {.tv_sec = sec, .tv_nsec = nsec};
 
-	if (!timespec_valid(&ts))
+	if (!timespec64_valid(&ts))
 		return -EINVAL;
 
 	/* Optimize for the zero timeout value here */
 	if (!sec && !nsec) {
 		to->tv_sec = to->tv_nsec = 0;
 	} else {
-		ktime_get_ts(to);
-		*to = timespec_add_safe(*to, ts);
+		ktime_get_ts64(to);
+		*to = timespec64_add_safe(*to, ts);
 	}
 	return 0;
 }
 
-static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
+static int poll_select_copy_remaining(struct timespec64 *end_time,
+				      void __user *p,
 				      int timeval, int ret)
 {
+	struct timespec64 rts64;
 	struct timespec rts;
 	struct timeval rtv;
 
@@ -302,16 +304,18 @@ static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
 	if (!end_time->tv_sec && !end_time->tv_nsec)
 		return ret;
 
-	ktime_get_ts(&rts);
-	rts = timespec_sub(*end_time, rts);
-	if (rts.tv_sec < 0)
-		rts.tv_sec = rts.tv_nsec = 0;
+	ktime_get_ts64(&rts64);
+	rts64 = timespec64_sub(*end_time, rts64);
+	if (rts64.tv_sec < 0)
+		rts64.tv_sec = rts64.tv_nsec = 0;
+
+	rts = timespec64_to_timespec(rts64);
 
 	if (timeval) {
 		if (sizeof(rtv) > sizeof(rtv.tv_sec) + sizeof(rtv.tv_usec))
 			memset(&rtv, 0, sizeof(rtv));
-		rtv.tv_sec = rts.tv_sec;
-		rtv.tv_usec = rts.tv_nsec / NSEC_PER_USEC;
+		rtv.tv_sec = rts64.tv_sec;
+		rtv.tv_usec = rts64.tv_nsec / NSEC_PER_USEC;
 
 		if (!copy_to_user(p, &rtv, sizeof(rtv)))
 			return ret;
@@ -396,7 +400,7 @@ static inline void wait_key_set(poll_table *wait, unsigned long in,
 		wait->_key |= POLLOUT_SET;
 }
 
-int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
+int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time)
 {
 	ktime_t expire, *to = NULL;
 	struct poll_wqueues table;
@@ -522,7 +526,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -545,7 +549,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
  * I'm trying ERESTARTNOHAND which restart only when you want to.
  */
 int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time)
+			   fd_set __user *exp, struct timespec64 *end_time)
 {
 	fd_set_bits fds;
 	void *bits;
@@ -622,7 +626,7 @@ out_nofds:
 SYSCALL_DEFINE5(select, int, n, fd_set __user *, inp, fd_set __user *, outp,
 		fd_set __user *, exp, struct timeval __user *, tvp)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	struct timeval tv;
 	int ret;
 
@@ -648,15 +652,17 @@ static long do_pselect(int n, fd_set __user *inp, fd_set __user *outp,
 		       const sigset_t __user *sigmask, size_t sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 ts64, end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
 		if (copy_from_user(&ts, tsp, sizeof(ts)))
 			return -EFAULT;
+		ts64 = timespec_to_timespec64(ts);
 
 		to = &end_time;
-		if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec))
+		if (poll_select_set_timeout(to, ts64.tv_sec, ts64.tv_nsec))
 			return -EINVAL;
 	}
 
@@ -779,7 +785,7 @@ static inline unsigned int do_pollfd(struct pollfd *pollfd, poll_table *pwait,
 }
 
 static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
-		   struct timespec *end_time)
+		   struct timespec64 *end_time)
 {
 	poll_table* pt = &wait->pt;
 	ktime_t expire, *to = NULL;
@@ -854,7 +860,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -868,7 +874,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 			sizeof(struct pollfd))
 
 int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
-		struct timespec *end_time)
+		struct timespec64 *end_time)
 {
 	struct poll_wqueues table;
  	int err = -EFAULT, fdcount, len, size;
@@ -936,7 +942,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 {
 	struct pollfd __user *ufds = restart_block->poll.ufds;
 	int nfds = restart_block->poll.nfds;
-	struct timespec *to = NULL, end_time;
+	struct timespec64 *to = NULL, end_time;
 	int ret;
 
 	if (restart_block->poll.has_timeout) {
@@ -957,7 +963,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds,
 		int, timeout_msecs)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (timeout_msecs >= 0) {
@@ -993,7 +999,8 @@ SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds, unsigned int, nfds,
 		size_t, sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
diff --git a/include/linux/poll.h b/include/linux/poll.h
index 9fb4f40..37b057b 100644
--- a/include/linux/poll.h
+++ b/include/linux/poll.h
@@ -96,7 +96,7 @@ extern void poll_initwait(struct poll_wqueues *pwq);
 extern void poll_freewait(struct poll_wqueues *pwq);
 extern int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
 				 ktime_t *expires, unsigned long slack);
-extern u64 select_estimate_accuracy(struct timespec *tv);
+extern u64 select_estimate_accuracy(struct timespec64 *tv);
 
 
 static inline int poll_schedule(struct poll_wqueues *pwq, int state)
@@ -153,12 +153,13 @@ void zero_fd_set(unsigned long nr, unsigned long *fdset)
 
 #define MAX_INT64_SECONDS (((s64)(~((u64)0)>>1)/HZ)-1)
 
-extern int do_select(int n, fd_set_bits *fds, struct timespec *end_time);
+extern int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time);
 extern int do_sys_poll(struct pollfd __user * ufds, unsigned int nfds,
-		       struct timespec *end_time);
+		       struct timespec64 *end_time);
 extern int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time);
+			   fd_set __user *exp, struct timespec64 *end_time);
 
-extern int poll_select_set_timeout(struct timespec *to, long sec, long nsec);
+extern int poll_select_set_timeout(struct timespec64 *to, time64_t sec,
+				   long nsec);
 
 #endif /* _LINUX_POLL_H */
diff --git a/net/socket.c b/net/socket.c
index 5dbb0bb..bdfe115 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -2171,7 +2171,8 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 	struct mmsghdr __user *entry;
 	struct compat_mmsghdr __user *compat_entry;
 	struct msghdr msg_sys;
-	struct timespec end_time;
+	struct timespec64 end_time;
+	struct timespec64 timeout64;
 
 	if (timeout &&
 	    poll_select_set_timeout(&end_time, timeout->tv_sec,
@@ -2223,8 +2224,9 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 			flags |= MSG_DONTWAIT;
 
 		if (timeout) {
-			ktime_get_ts(timeout);
-			*timeout = timespec_sub(end_time, *timeout);
+			ktime_get_ts64(&timeout64);
+			*timeout = timespec64_to_timespec(
+					timespec64_sub(end_time, timeout64));
 			if (timeout->tv_sec < 0) {
 				timeout->tv_sec = timeout->tv_nsec = 0;
 				break;
-- 
1.9.1

[RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Deepa Dinamani]

struct timespec is not y2038 safe.
Even though timespec might be sufficient to represent
timeouts, use struct timespec64 here as the plan is to
get rid of all timespec reference in the kernel.

The patch transitions the common functions:
poll_select_set_timeout() and select_estimate_accuracy()
to use timespec64. And, all the syscalls that use these
functions are transitioned in the same patch.

The restart block parameters for poll uses monotonic time.
Use timespec64 here as well to assign timeout value. This
parameter in the restart block need not change because
this only holds the monotonic timestamp at which timeout
should occur. And, unsigned long data type should be big
enough for this timestamp.

The system call interfaces will be handled in a separate
series.

Compat interfaces need not change as timespec64 is an
alias to struct timespec on a 64 bit system.

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
---
Resending to include John and Thomas on this patch as well.
This is to include this patch also in John's tree.
This will let all 3 patches in the series to merged through the same tree.
 
 fs/eventpoll.c       | 12 +++++-----
 fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
 include/linux/poll.h | 11 +++++----
 net/socket.c         |  8 ++++---
 4 files changed, 54 insertions(+), 44 deletions(-)

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 8a74a2a..10db912 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -1583,15 +1583,15 @@ static int ep_send_events(struct eventpoll *ep,
 	return ep_scan_ready_list(ep, ep_send_events_proc, &esed, 0, false);
 }
 
-static inline struct timespec ep_set_mstimeout(long ms)
+static inline struct timespec64 ep_set_mstimeout(long ms)
 {
-	struct timespec now, ts = {
+	struct timespec64 now, ts = {
 		.tv_sec = ms / MSEC_PER_SEC,
 		.tv_nsec = NSEC_PER_MSEC * (ms % MSEC_PER_SEC),
 	};
 
-	ktime_get_ts(&now);
-	return timespec_add_safe(now, ts);
+	ktime_get_ts64(&now);
+	return timespec64_add_safe(now, ts);
 }
 
 /**
@@ -1621,11 +1621,11 @@ static int ep_poll(struct eventpoll *ep, struct epoll_event __user *events,
 	ktime_t expires, *to = NULL;
 
 	if (timeout > 0) {
-		struct timespec end_time = ep_set_mstimeout(timeout);
+		struct timespec64 end_time = ep_set_mstimeout(timeout);
 
 		slack = select_estimate_accuracy(&end_time);
 		to = &expires;
-		*to = timespec_to_ktime(end_time);
+		*to = timespec64_to_ktime(end_time);
 	} else if (timeout == 0) {
 		/*
 		 * Avoid the unnecessary trip to the wait queue loop, if the
diff --git a/fs/select.c b/fs/select.c
index 8692939..8ed9da5 100644
--- a/fs/select.c
+++ b/fs/select.c
@@ -47,7 +47,7 @@
 
 #define MAX_SLACK	(100 * NSEC_PER_MSEC)
 
-static long __estimate_accuracy(struct timespec *tv)
+static long __estimate_accuracy(struct timespec64 *tv)
 {
 	long slack;
 	int divfactor = 1000;
@@ -70,10 +70,10 @@ static long __estimate_accuracy(struct timespec *tv)
 	return slack;
 }
 
-u64 select_estimate_accuracy(struct timespec *tv)
+u64 select_estimate_accuracy(struct timespec64 *tv)
 {
 	u64 ret;
-	struct timespec now;
+	struct timespec64 now;
 
 	/*
 	 * Realtime tasks get a slack of 0 for obvious reasons.
@@ -82,8 +82,8 @@ u64 select_estimate_accuracy(struct timespec *tv)
 	if (rt_task(current))
 		return 0;
 
-	ktime_get_ts(&now);
-	now = timespec_sub(*tv, now);
+	ktime_get_ts64(&now);
+	now = timespec64_sub(*tv, now);
 	ret = __estimate_accuracy(&now);
 	if (ret < current->timer_slack_ns)
 		return current->timer_slack_ns;
@@ -260,7 +260,7 @@ EXPORT_SYMBOL(poll_schedule_timeout);
 
 /**
  * poll_select_set_timeout - helper function to setup the timeout value
- * @to:		pointer to timespec variable for the final timeout
+ * @to:		pointer to timespec64 variable for the final timeout
  * @sec:	seconds (from user space)
  * @nsec:	nanoseconds (from user space)
  *
@@ -269,26 +269,28 @@ EXPORT_SYMBOL(poll_schedule_timeout);
  *
  * Returns -EINVAL if sec/nsec are not normalized. Otherwise 0.
  */
-int poll_select_set_timeout(struct timespec *to, long sec, long nsec)
+int poll_select_set_timeout(struct timespec64 *to, time64_t sec, long nsec)
 {
-	struct timespec ts = {.tv_sec = sec, .tv_nsec = nsec};
+	struct timespec64 ts = {.tv_sec = sec, .tv_nsec = nsec};
 
-	if (!timespec_valid(&ts))
+	if (!timespec64_valid(&ts))
 		return -EINVAL;
 
 	/* Optimize for the zero timeout value here */
 	if (!sec && !nsec) {
 		to->tv_sec = to->tv_nsec = 0;
 	} else {
-		ktime_get_ts(to);
-		*to = timespec_add_safe(*to, ts);
+		ktime_get_ts64(to);
+		*to = timespec64_add_safe(*to, ts);
 	}
 	return 0;
 }
 
-static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
+static int poll_select_copy_remaining(struct timespec64 *end_time,
+				      void __user *p,
 				      int timeval, int ret)
 {
+	struct timespec64 rts64;
 	struct timespec rts;
 	struct timeval rtv;
 
@@ -302,16 +304,18 @@ static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
 	if (!end_time->tv_sec && !end_time->tv_nsec)
 		return ret;
 
-	ktime_get_ts(&rts);
-	rts = timespec_sub(*end_time, rts);
-	if (rts.tv_sec < 0)
-		rts.tv_sec = rts.tv_nsec = 0;
+	ktime_get_ts64(&rts64);
+	rts64 = timespec64_sub(*end_time, rts64);
+	if (rts64.tv_sec < 0)
+		rts64.tv_sec = rts64.tv_nsec = 0;
+
+	rts = timespec64_to_timespec(rts64);
 
 	if (timeval) {
 		if (sizeof(rtv) > sizeof(rtv.tv_sec) + sizeof(rtv.tv_usec))
 			memset(&rtv, 0, sizeof(rtv));
-		rtv.tv_sec = rts.tv_sec;
-		rtv.tv_usec = rts.tv_nsec / NSEC_PER_USEC;
+		rtv.tv_sec = rts64.tv_sec;
+		rtv.tv_usec = rts64.tv_nsec / NSEC_PER_USEC;
 
 		if (!copy_to_user(p, &rtv, sizeof(rtv)))
 			return ret;
@@ -396,7 +400,7 @@ static inline void wait_key_set(poll_table *wait, unsigned long in,
 		wait->_key |= POLLOUT_SET;
 }
 
-int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
+int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time)
 {
 	ktime_t expire, *to = NULL;
 	struct poll_wqueues table;
@@ -522,7 +526,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -545,7 +549,7 @@ int do_select(int n, fd_set_bits *fds, struct timespec *end_time)
  * I'm trying ERESTARTNOHAND which restart only when you want to.
  */
 int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time)
+			   fd_set __user *exp, struct timespec64 *end_time)
 {
 	fd_set_bits fds;
 	void *bits;
@@ -622,7 +626,7 @@ out_nofds:
 SYSCALL_DEFINE5(select, int, n, fd_set __user *, inp, fd_set __user *, outp,
 		fd_set __user *, exp, struct timeval __user *, tvp)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	struct timeval tv;
 	int ret;
 
@@ -648,15 +652,17 @@ static long do_pselect(int n, fd_set __user *inp, fd_set __user *outp,
 		       const sigset_t __user *sigmask, size_t sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 ts64, end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
 		if (copy_from_user(&ts, tsp, sizeof(ts)))
 			return -EFAULT;
+		ts64 = timespec_to_timespec64(ts);
 
 		to = &end_time;
-		if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec))
+		if (poll_select_set_timeout(to, ts64.tv_sec, ts64.tv_nsec))
 			return -EINVAL;
 	}
 
@@ -779,7 +785,7 @@ static inline unsigned int do_pollfd(struct pollfd *pollfd, poll_table *pwait,
 }
 
 static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
-		   struct timespec *end_time)
+		   struct timespec64 *end_time)
 {
 	poll_table* pt = &wait->pt;
 	ktime_t expire, *to = NULL;
@@ -854,7 +860,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 		 * pointer to the expiry value.
 		 */
 		if (end_time && !to) {
-			expire = timespec_to_ktime(*end_time);
+			expire = timespec64_to_ktime(*end_time);
 			to = &expire;
 		}
 
@@ -868,7 +874,7 @@ static int do_poll(struct poll_list *list, struct poll_wqueues *wait,
 			sizeof(struct pollfd))
 
 int do_sys_poll(struct pollfd __user *ufds, unsigned int nfds,
-		struct timespec *end_time)
+		struct timespec64 *end_time)
 {
 	struct poll_wqueues table;
  	int err = -EFAULT, fdcount, len, size;
@@ -936,7 +942,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 {
 	struct pollfd __user *ufds = restart_block->poll.ufds;
 	int nfds = restart_block->poll.nfds;
-	struct timespec *to = NULL, end_time;
+	struct timespec64 *to = NULL, end_time;
 	int ret;
 
 	if (restart_block->poll.has_timeout) {
@@ -957,7 +963,7 @@ static long do_restart_poll(struct restart_block *restart_block)
 SYSCALL_DEFINE3(poll, struct pollfd __user *, ufds, unsigned int, nfds,
 		int, timeout_msecs)
 {
-	struct timespec end_time, *to = NULL;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (timeout_msecs >= 0) {
@@ -993,7 +999,8 @@ SYSCALL_DEFINE5(ppoll, struct pollfd __user *, ufds, unsigned int, nfds,
 		size_t, sigsetsize)
 {
 	sigset_t ksigmask, sigsaved;
-	struct timespec ts, end_time, *to = NULL;
+	struct timespec ts;
+	struct timespec64 end_time, *to = NULL;
 	int ret;
 
 	if (tsp) {
diff --git a/include/linux/poll.h b/include/linux/poll.h
index 9fb4f40..37b057b 100644
--- a/include/linux/poll.h
+++ b/include/linux/poll.h
@@ -96,7 +96,7 @@ extern void poll_initwait(struct poll_wqueues *pwq);
 extern void poll_freewait(struct poll_wqueues *pwq);
 extern int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
 				 ktime_t *expires, unsigned long slack);
-extern u64 select_estimate_accuracy(struct timespec *tv);
+extern u64 select_estimate_accuracy(struct timespec64 *tv);
 
 
 static inline int poll_schedule(struct poll_wqueues *pwq, int state)
@@ -153,12 +153,13 @@ void zero_fd_set(unsigned long nr, unsigned long *fdset)
 
 #define MAX_INT64_SECONDS (((s64)(~((u64)0)>>1)/HZ)-1)
 
-extern int do_select(int n, fd_set_bits *fds, struct timespec *end_time);
+extern int do_select(int n, fd_set_bits *fds, struct timespec64 *end_time);
 extern int do_sys_poll(struct pollfd __user * ufds, unsigned int nfds,
-		       struct timespec *end_time);
+		       struct timespec64 *end_time);
 extern int core_sys_select(int n, fd_set __user *inp, fd_set __user *outp,
-			   fd_set __user *exp, struct timespec *end_time);
+			   fd_set __user *exp, struct timespec64 *end_time);
 
-extern int poll_select_set_timeout(struct timespec *to, long sec, long nsec);
+extern int poll_select_set_timeout(struct timespec64 *to, time64_t sec,
+				   long nsec);
 
 #endif /* _LINUX_POLL_H */
diff --git a/net/socket.c b/net/socket.c
index 5dbb0bb..bdfe115 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -2171,7 +2171,8 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 	struct mmsghdr __user *entry;
 	struct compat_mmsghdr __user *compat_entry;
 	struct msghdr msg_sys;
-	struct timespec end_time;
+	struct timespec64 end_time;
+	struct timespec64 timeout64;
 
 	if (timeout &&
 	    poll_select_set_timeout(&end_time, timeout->tv_sec,
@@ -2223,8 +2224,9 @@ int __sys_recvmmsg(int fd, struct mmsghdr __user *mmsg, unsigned int vlen,
 			flags |= MSG_DONTWAIT;
 
 		if (timeout) {
-			ktime_get_ts(timeout);
-			*timeout = timespec_sub(end_time, *timeout);
+			ktime_get_ts64(&timeout64);
+			*timeout = timespec64_to_timespec(
+					timespec64_sub(end_time, timeout64));
 			if (timeout->tv_sec < 0) {
 				timeout->tv_sec = timeout->tv_nsec = 0;
 				break;
-- 
1.9.1

_______________________________________________
Y2038 mailing list
Y2038@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/y2038

Re: [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[John Stultz]

On Wed, May 4, 2016 at 12:24 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> struct timespec is not y2038 safe.
> Even though timespec might be sufficient to represent
> timeouts, use struct timespec64 here as the plan is to
> get rid of all timespec reference in the kernel.
>
> The patch transitions the common functions:
> poll_select_set_timeout() and select_estimate_accuracy()
> to use timespec64. And, all the syscalls that use these
> functions are transitioned in the same patch.
>
> The restart block parameters for poll uses monotonic time.
> Use timespec64 here as well to assign timeout value. This
> parameter in the restart block need not change because
> this only holds the monotonic timestamp at which timeout
> should occur. And, unsigned long data type should be big
> enough for this timestamp.
>
> The system call interfaces will be handled in a separate
> series.
>
> Compat interfaces need not change as timespec64 is an
> alias to struct timespec on a 64 bit system.
>
> Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
> Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: netdev@vger.kernel.org
> ---
> Resending to include John and Thomas on this patch as well.
> This is to include this patch also in John's tree.
> This will let all 3 patches in the series to merged through the same tree.
>
>  fs/eventpoll.c       | 12 +++++-----
>  fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
>  include/linux/poll.h | 11 +++++----
>  net/socket.c         |  8 ++++---
>  4 files changed, 54 insertions(+), 44 deletions(-)


So with the other two patches, I'm more comfortable queueing and
sending through to Thomas.

But I'm less comfortable making the call on this one. It looks
relatively straight forward, but it would be good to have maintainer
acks before I add it to my tree.

thanks
-john

Re: [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[John Stultz]

On Wed, May 4, 2016 at 12:24 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> struct timespec is not y2038 safe.
> Even though timespec might be sufficient to represent
> timeouts, use struct timespec64 here as the plan is to
> get rid of all timespec reference in the kernel.
>
> The patch transitions the common functions:
> poll_select_set_timeout() and select_estimate_accuracy()
> to use timespec64. And, all the syscalls that use these
> functions are transitioned in the same patch.
>
> The restart block parameters for poll uses monotonic time.
> Use timespec64 here as well to assign timeout value. This
> parameter in the restart block need not change because
> this only holds the monotonic timestamp at which timeout
> should occur. And, unsigned long data type should be big
> enough for this timestamp.
>
> The system call interfaces will be handled in a separate
> series.
>
> Compat interfaces need not change as timespec64 is an
> alias to struct timespec on a 64 bit system.
>
> Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
> Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> Cc: "David S. Miller" <davem@davemloft.net>
> Cc: netdev@vger.kernel.org
> ---
> Resending to include John and Thomas on this patch as well.
> This is to include this patch also in John's tree.
> This will let all 3 patches in the series to merged through the same tree.
>
>  fs/eventpoll.c       | 12 +++++-----
>  fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
>  include/linux/poll.h | 11 +++++----
>  net/socket.c         |  8 ++++---
>  4 files changed, 54 insertions(+), 44 deletions(-)


So with the other two patches, I'm more comfortable queueing and
sending through to Thomas.

But I'm less comfortable making the call on this one. It looks
relatively straight forward, but it would be good to have maintainer
acks before I add it to my tree.

thanks
-john
_______________________________________________
Y2038 mailing list
Y2038@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/y2038

Re: [Y2038] [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Arnd Bergmann]

On Wednesday 04 May 2016 13:04:37 John Stultz wrote:
> On Wed, May 4, 2016 at 12:24 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> > struct timespec is not y2038 safe.
> > Even though timespec might be sufficient to represent
> > timeouts, use struct timespec64 here as the plan is to
> > get rid of all timespec reference in the kernel.
> >
> > The patch transitions the common functions:
> > poll_select_set_timeout() and select_estimate_accuracy()
> > to use timespec64. And, all the syscalls that use these
> > functions are transitioned in the same patch.
> >
> > The restart block parameters for poll uses monotonic time.
> > Use timespec64 here as well to assign timeout value. This
> > parameter in the restart block need not change because
> > this only holds the monotonic timestamp at which timeout
> > should occur. And, unsigned long data type should be big
> > enough for this timestamp.
> >
> > The system call interfaces will be handled in a separate
> > series.
> >
> > Compat interfaces need not change as timespec64 is an
> > alias to struct timespec on a 64 bit system.
> >
> > Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
> > Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> > Cc: "David S. Miller" <davem@davemloft.net>
> > Cc: netdev@vger.kernel.org
> > ---
> > Resending to include John and Thomas on this patch as well.
> > This is to include this patch also in John's tree.
> > This will let all 3 patches in the series to merged through the same tree.
> >
> >  fs/eventpoll.c       | 12 +++++-----
> >  fs/select.c          | 67 +++++++++++++++++++++++++++++-----------------------
> >  include/linux/poll.h | 11 +++++----
> >  net/socket.c         |  8 ++++---
> >  4 files changed, 54 insertions(+), 44 deletions(-)
> 
> 
> So with the other two patches, I'm more comfortable queueing and
> sending through to Thomas.

Note that of course patch 3 depends on patch 2, otherwise it would
be a simple rename.

> But I'm less comfortable making the call on this one. It looks
> relatively straight forward, but it would be good to have maintainer
> acks before I add it to my tree.

Agreed. Feel free to add my

Reviewed-by: Arnd Bergmann <arnd@arndb.de>

at least (whoever picks it up). The file hasn't changed much in the past
decade, these are all the people who did more than 1 patch for fs/select.c
in git history:

      7 Al Viro
      6 Eliezer Tamir
      6 Arjan van de Ven
      4 Andrew Morton
      3 Linus Torvalds
      3 Heiko Carstens
      2 Vadim Lobanov
      2 Roland McGrath
      2 Oleg Nesterov
      2 Jiri Slaby
      2 Guillaume Knispel
      2 Eric Dumazet
      2 Dipankar Sarma
      2 Alexey Dobriyan

adding a few more to Cc, maybe someone else finds the time to take a
second look.

	Arnd

Re: [Y2038] [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Andrew Morton]

On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:

> > But I'm less comfortable making the call on this one. It looks
> > relatively straight forward, but it would be good to have maintainer
> > acks before I add it to my tree.
> 
> Agreed. Feel free to add my
> 
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> 
> at least (whoever picks it up).

In reply to [1/3] John said

: Looks ok at the first glance. I've queued these up for testing,
: however I only got #1 and #3 of the set. Are you hoping these two
: patches will go through tip/timers/core or are you looking for acks so
: they can go via another tree?

However none of the patches are in linux-next.

John had qualms about [2/3], but it looks like a straightforward
substitution in areas which will get plenty of testing

I'll grab the patches for now to get them some external testing.

Re: [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[Andrew Morton]

On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:

> > But I'm less comfortable making the call on this one. It looks
> > relatively straight forward, but it would be good to have maintainer
> > acks before I add it to my tree.
> 
> Agreed. Feel free to add my
> 
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> 
> at least (whoever picks it up).

In reply to [1/3] John said

: Looks ok at the first glance. I've queued these up for testing,
: however I only got #1 and #3 of the set. Are you hoping these two
: patches will go through tip/timers/core or are you looking for acks so
: they can go via another tree?

However none of the patches are in linux-next.

John had qualms about [2/3], but it looks like a straightforward
substitution in areas which will get plenty of testing

I'll grab the patches for now to get them some external testing.
_______________________________________________
Y2038 mailing list
Y2038@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/y2038

Re: [Y2038] [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[John Stultz]

On Wed, May 4, 2016 at 4:51 PM, Andrew Morton <akpm@linux-foundation.org> wrote:
> On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:
>
>> > But I'm less comfortable making the call on this one. It looks
>> > relatively straight forward, but it would be good to have maintainer
>> > acks before I add it to my tree.
>>
>> Agreed. Feel free to add my
>>
>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>>
>> at least (whoever picks it up).
>
> In reply to [1/3] John said
>
> : Looks ok at the first glance. I've queued these up for testing,
> : however I only got #1 and #3 of the set. Are you hoping these two
> : patches will go through tip/timers/core or are you looking for acks so
> : they can go via another tree?
>
> However none of the patches are in linux-next.
>
> John had qualms about [2/3], but it looks like a straightforward
> substitution in areas which will get plenty of testing

Yea. My main concern is just not stepping on any other maintainers toes.

> I'll grab the patches for now to get them some external testing.

I'd be just as happy if the set went through you Andrew.

For the set: Acked-by: John Stultz <john.stultz@linaro.org>

thanks
-john

Re: [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[John Stultz]

On Wed, May 4, 2016 at 4:51 PM, Andrew Morton <akpm@linux-foundation.org> wrote:
> On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:
>
>> > But I'm less comfortable making the call on this one. It looks
>> > relatively straight forward, but it would be good to have maintainer
>> > acks before I add it to my tree.
>>
>> Agreed. Feel free to add my
>>
>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>>
>> at least (whoever picks it up).
>
> In reply to [1/3] John said
>
> : Looks ok at the first glance. I've queued these up for testing,
> : however I only got #1 and #3 of the set. Are you hoping these two
> : patches will go through tip/timers/core or are you looking for acks so
> : they can go via another tree?
>
> However none of the patches are in linux-next.
>
> John had qualms about [2/3], but it looks like a straightforward
> substitution in areas which will get plenty of testing

Yea. My main concern is just not stepping on any other maintainers toes.

> I'll grab the patches for now to get them some external testing.

I'd be just as happy if the set went through you Andrew.

For the set: Acked-by: John Stultz <john.stultz@linaro.org>

thanks
-john
_______________________________________________
Y2038 mailing list
Y2038@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/y2038

Re: [Y2038] [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[David Miller]

From: John Stultz <john.stultz@linaro.org>
Date: Wed, 4 May 2016 17:01:24 -0700

> On Wed, May 4, 2016 at 4:51 PM, Andrew Morton <akpm@linux-foundation.org> wrote:
>> On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:
>>
>>> > But I'm less comfortable making the call on this one. It looks
>>> > relatively straight forward, but it would be good to have maintainer
>>> > acks before I add it to my tree.
>>>
>>> Agreed. Feel free to add my
>>>
>>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>>>
>>> at least (whoever picks it up).
>>
>> In reply to [1/3] John said
>>
>> : Looks ok at the first glance. I've queued these up for testing,
>> : however I only got #1 and #3 of the set. Are you hoping these two
>> : patches will go through tip/timers/core or are you looking for acks so
>> : they can go via another tree?
>>
>> However none of the patches are in linux-next.
>>
>> John had qualms about [2/3], but it looks like a straightforward
>> substitution in areas which will get plenty of testing
> 
> Yea. My main concern is just not stepping on any other maintainers toes.

The networking changes look fine to me:

Acked-by: David S. Miller <davem@davemloft.net>

Re: [RESEND PATCH 2/3] fs: poll/select/recvmmsg: use timespec64 for timeout events

[David Miller]

From: John Stultz <john.stultz@linaro.org>
Date: Wed, 4 May 2016 17:01:24 -0700

> On Wed, May 4, 2016 at 4:51 PM, Andrew Morton <akpm@linux-foundation.org> wrote:
>> On Wed, 04 May 2016 23:08:11 +0200 Arnd Bergmann <arnd@arndb.de> wrote:
>>
>>> > But I'm less comfortable making the call on this one. It looks
>>> > relatively straight forward, but it would be good to have maintainer
>>> > acks before I add it to my tree.
>>>
>>> Agreed. Feel free to add my
>>>
>>> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
>>>
>>> at least (whoever picks it up).
>>
>> In reply to [1/3] John said
>>
>> : Looks ok at the first glance. I've queued these up for testing,
>> : however I only got #1 and #3 of the set. Are you hoping these two
>> : patches will go through tip/timers/core or are you looking for acks so
>> : they can go via another tree?
>>
>> However none of the patches are in linux-next.
>>
>> John had qualms about [2/3], but it looks like a straightforward
>> substitution in areas which will get plenty of testing
> 
> Yea. My main concern is just not stepping on any other maintainers toes.

The networking changes look fine to me:

Acked-by: David S. Miller <davem@davemloft.net>
_______________________________________________
Y2038 mailing list
Y2038@lists.linaro.org
https://lists.linaro.org/mailman/listinfo/y2038