[PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Deepa Dinamani]

The series is a preparation series for individual architectures
to use 64 bit time_t syscalls in compat and 32 bit emulation modes.

This is a follow up to the series Arnd Bergmann posted:
https://sourceware.org/ml/libc-alpha/2015-05/msg00070.html [1]

Big picture is as per the lwn article:
https://lwn.net/Articles/643234/ [2]

The series is directed at converting posix clock syscalls:
clock_gettime, clock_settime, clock_getres and clock_nanosleep
to use a new data structure __kernel_timespec at syscall boundaries.
__kernel_timespec maintains 64 bit time_t across all execution modes.

vdso will be handled as part of each architecture when they enable
support for 64 bit time_t.

The compat syscalls are repurposed to provide backward compatibility
by using them as native syscalls as well for 32 bit architectures.
They will continue to use timespec at syscall boundaries.

CONFIG_64_BIT_TIME controls whether the syscalls use __kernel_timespec
or timespec at syscall boundaries.

The series does the following:
1. Enable compat syscalls on 32 bit architectures.
2. Add a new __kernel_timespec type to be used as the data structure
   for all the new syscalls.
3. Add new config CONFIG_64BIT_TIME(intead of the CONFIG_COMPAT_TIME in
   [1] and [2] to switch to new definition of __kernel_timespec. It is
   the same as struct timespec otherwise.
4. Add new CONFIG_32BIT_TIME to conditionally compile compat syscalls.

* Changes since v1:
 * Introduce CONFIG_32BIT_TIME
 * Fixed zeroing out of higher order bits of tv_nsec
 * Included Arnd's changes to fix up use of compat headers

I decided against using LEGACY_TIME_SYSCALLS to conditionally compile
legacy time syscalls such as sys_nanosleep because this will need to
enclose compat_sys_nanosleep as well. So, defining it as 

config LEGACY_TIME_SYSCALLS
     def_bool 64BIT || !64BIT_TIME

will not include compat_sys_nanosleep. We will instead need a new config to
exclusively mark legacy syscalls.

Deepa Dinamani (10):
  compat: Make compat helpers independent of CONFIG_COMPAT
  include: Move compat_timespec/ timeval to compat_time.h
  compat: enable compat_get/put_timespec64 always
  arch: introduce CONFIG_64BIT_TIME
  arch: Introduce CONFIG_COMPAT_32BIT_TIME
  posix-clocks: Make compat syscalls depend on CONFIG_COMPAT_32BIT_TIME
  include: Add new y2038 safe __kernel_timespec
  fix get_timespec64() for y2038 safe compat interfaces
  change time types to new y2038 safe __kernel_* types
  nanosleep: change time types to safe __kernel_* types

 arch/Kconfig                           | 18 +++++++
 arch/arm64/include/asm/compat.h        | 11 ----
 arch/arm64/include/asm/stat.h          |  1 +
 arch/arm64/kernel/hw_breakpoint.c      |  1 -
 arch/arm64/kernel/perf_regs.c          |  2 +-
 arch/arm64/kernel/process.c            |  1 -
 arch/mips/include/asm/compat.h         | 11 ----
 arch/mips/kernel/signal32.c            |  2 +-
 arch/parisc/include/asm/compat.h       | 11 ----
 arch/powerpc/include/asm/compat.h      | 11 ----
 arch/powerpc/kernel/asm-offsets.c      |  2 +-
 arch/powerpc/oprofile/backtrace.c      |  2 +-
 arch/s390/hypfs/hypfs_sprp.c           |  1 -
 arch/s390/include/asm/compat.h         | 11 ----
 arch/s390/include/asm/elf.h            |  3 +-
 arch/s390/kvm/priv.c                   |  1 -
 arch/s390/pci/pci_clp.c                |  1 -
 arch/sparc/include/asm/compat.h        | 11 ----
 arch/tile/include/asm/compat.h         | 11 ----
 arch/x86/events/core.c                 |  2 +-
 arch/x86/include/asm/compat.h          | 11 ----
 arch/x86/include/asm/ftrace.h          |  2 +-
 arch/x86/include/asm/sys_ia32.h        |  2 +-
 arch/x86/kernel/sys_x86_64.c           |  2 +-
 drivers/s390/block/dasd_ioctl.c        |  1 -
 drivers/s390/char/fs3270.c             |  1 -
 drivers/s390/char/sclp_ctl.c           |  1 -
 drivers/s390/char/vmcp.c               |  1 -
 drivers/s390/cio/chsc_sch.c            |  1 -
 drivers/s390/net/qeth_core_main.c      |  2 +-
 drivers/staging/pi433/pi433_if.c       |  2 +-
 include/linux/compat.h                 | 11 ++--
 include/linux/compat_time.h            | 23 +++++++++
 include/linux/restart_block.h          |  7 +--
 include/linux/syscalls.h               | 12 ++---
 include/linux/time.h                   |  4 +-
 include/linux/time64.h                 | 10 +++-
 include/uapi/asm-generic/posix_types.h |  1 +
 include/uapi/linux/time.h              |  7 +++
 kernel/Makefile                        |  2 +-
 kernel/compat.c                        | 92 ++++++++++++++++++----------------
 kernel/time/hrtimer.c                  | 10 ++--
 kernel/time/posix-stubs.c              | 12 +++--
 kernel/time/posix-timers.c             | 24 ++++++---
 kernel/time/time.c                     | 10 +++-
 45 files changed, 175 insertions(+), 190 deletions(-)
 create mode 100644 include/linux/compat_time.h

base-commit: b0a84f19a5161418d4360cd57603e94ed489915e
-- 
2.14.1

Cc: acme@kernel.org
Cc: benh@kernel.crashing.org
Cc: borntraeger@de.ibm.com
Cc: catalin.marinas@arm.com
Cc: cmetcalf@mellanox.com
Cc: cohuck@redhat.com
Cc: davem@davemloft.net
Cc: deller@gmx.de
Cc: devel@driverdev.osuosl.org
Cc: gerald.schaefer@de.ibm.com
Cc: gregkh@linuxfoundation.org
Cc: heiko.carstens@de.ibm.com
Cc: hoeppner@linux.vnet.ibm.com
Cc: hpa@zytor.com
Cc: jejb@parisc-linux.org
Cc: jwi@linux.vnet.ibm.com
Cc: linux-api@vger.kernel.org
Cc: linux-arch@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-mips@linux-mips.org
Cc: linux-parisc@vger.kernel.org
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-s390@vger.kernel.org
Cc: mark.rutland@arm.com
Cc: mingo@redhat.com
Cc: mpe@ellerman.id.au
Cc: oberpar@linux.vnet.ibm.com
Cc: oprofile-list@lists.sf.net
Cc: paulus@samba.org
Cc: peterz@infradead.org
Cc: ralf@linux-mips.org
Cc: rostedt@goodmis.org
Cc: rric@kernel.org
Cc: schwidefsky@de.ibm.com
Cc: sebott@linux.vnet.ibm.com
Cc: sparclinux@vger.kernel.org
Cc: sth@linux.vnet.ibm.com
Cc: ubraun@linux.vnet.ibm.com
Cc: will.deacon@arm.com
Cc: x86@kernel.org

[PATCH v2 09/10] change time types to new y2038 safe __kernel_* types

[Deepa Dinamani]

Change over clock_settime, clock_gettime and clock_getres
syscalls to use __kernel_timespec times. This will enable
changing over of these syscalls to use new y2038 safe syscalls
when the architectures define the CONFIG_64BIT_TIME.

Cc: linux-api@vger.kernel.org
Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
---
 include/linux/syscalls.h   | 6 +++---
 kernel/time/posix-stubs.c  | 6 +++---
 kernel/time/posix-timers.c | 6 +++---
 3 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index a78186d826d7..7ac1bb9ea7da 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -309,13 +309,13 @@ asmlinkage long sys_timer_settime(timer_t timer_id, int flags,
 				struct itimerspec __user *old_setting);
 asmlinkage long sys_timer_delete(timer_t timer_id);
 asmlinkage long sys_clock_settime(clockid_t which_clock,
-				const struct timespec __user *tp);
+				const struct __kernel_timespec __user *tp);
 asmlinkage long sys_clock_gettime(clockid_t which_clock,
-				struct timespec __user *tp);
+				struct __kernel_timespec __user *tp);
 asmlinkage long sys_clock_adjtime(clockid_t which_clock,
 				struct timex __user *tx);
 asmlinkage long sys_clock_getres(clockid_t which_clock,
-				struct timespec __user *tp);
+				struct __kernel_timespec __user *tp);
 asmlinkage long sys_clock_nanosleep(clockid_t which_clock, int flags,
 				const struct timespec __user *rqtp,
 				struct timespec __user *rmtp);
diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c
index 0645cfa93609..28c22f54d0f2 100644
--- a/kernel/time/posix-stubs.c
+++ b/kernel/time/posix-stubs.c
@@ -49,7 +49,7 @@ SYS_NI(alarm);
  */
 
 SYSCALL_DEFINE2(clock_settime, const clockid_t, which_clock,
-		const struct timespec __user *, tp)
+		const struct __kernel_timespec __user *, tp)
 {
 	struct timespec64 new_tp;
 
@@ -80,7 +80,7 @@ int do_clock_gettime(clockid_t which_clock, struct timespec64 *tp)
 	return 0;
 }
 SYSCALL_DEFINE2(clock_gettime, const clockid_t, which_clock,
-		struct timespec __user *, tp)
+		struct __kernel_timespec __user *, tp)
 {
 	int ret;
 	struct timespec64 kernel_tp;
@@ -94,7 +94,7 @@ SYSCALL_DEFINE2(clock_gettime, const clockid_t, which_clock,
 	return 0;
 }
 
-SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock, struct timespec __user *, tp)
+SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock, struct __kernel_timespec __user *, tp)
 {
 	struct timespec64 rtn_tp = {
 		.tv_sec = 0,
diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c
index 67ed5966eccd..d46ee8264fd3 100644
--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -1034,7 +1034,7 @@ void exit_itimers(struct signal_struct *sig)
 }
 
 SYSCALL_DEFINE2(clock_settime, const clockid_t, which_clock,
-		const struct timespec __user *, tp)
+		const struct __kernel_timespec __user *, tp)
 {
 	const struct k_clock *kc = clockid_to_kclock(which_clock);
 	struct timespec64 new_tp;
@@ -1049,7 +1049,7 @@ SYSCALL_DEFINE2(clock_settime, const clockid_t, which_clock,
 }
 
 SYSCALL_DEFINE2(clock_gettime, const clockid_t, which_clock,
-		struct timespec __user *,tp)
+		struct __kernel_timespec __user *, tp)
 {
 	const struct k_clock *kc = clockid_to_kclock(which_clock);
 	struct timespec64 kernel_tp;
@@ -1090,7 +1090,7 @@ SYSCALL_DEFINE2(clock_adjtime, const clockid_t, which_clock,
 }
 
 SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock,
-		struct timespec __user *, tp)
+		struct __kernel_timespec __user *, tp)
 {
 	const struct k_clock *kc = clockid_to_kclock(which_clock);
 	struct timespec64 rtn_tp;
-- 
2.14.1

[PATCH v2 10/10] nanosleep: change time types to safe __kernel_* types

[Deepa Dinamani]

Change over clock_nanosleep syscalls to use y2038 safe
__kernel_timespec times. This will enable changing over
of these syscalls to use new y2038 safe syscalls when
the architectures define the CONFIG_64BIT_TIME.

Note that nanosleep syscall is deprecated and does not have a
plan for making it y2038 safe. But, the syscall should work as
before on 64 bit machines and on 32 bit machines, the syscall
works correctly until y2038 as before using the existing compat
syscall version. There is no new syscall for supporting 64 bit
time_t on 32 bit architectures.

Cc: linux-api-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Signed-off-by: Deepa Dinamani <deepa.kernel-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
---
 include/linux/restart_block.h | 7 ++-----
 include/linux/syscalls.h      | 6 +++---
 kernel/time/hrtimer.c         | 8 ++++++--
 kernel/time/posix-stubs.c     | 4 ++--
 kernel/time/posix-timers.c    | 4 ++--
 5 files changed, 15 insertions(+), 14 deletions(-)

diff --git a/include/linux/restart_block.h b/include/linux/restart_block.h
index bcfdb918cd81..5d83d0c1d06c 100644
--- a/include/linux/restart_block.h
+++ b/include/linux/restart_block.h
@@ -7,6 +7,7 @@
 
 #include <linux/compiler.h>
 #include <linux/types.h>
+#include <linux/time64.h>
 
 struct timespec;
 struct compat_timespec;
@@ -15,9 +16,7 @@ struct pollfd;
 enum timespec_type {
 	TT_NONE		= 0,
 	TT_NATIVE	= 1,
-#ifdef CONFIG_COMPAT
 	TT_COMPAT	= 2,
-#endif
 };
 
 /*
@@ -40,10 +39,8 @@ struct restart_block {
 			clockid_t clockid;
 			enum timespec_type type;
 			union {
-				struct timespec __user *rmtp;
-#ifdef CONFIG_COMPAT
+				struct __kernel_timespec __user *rmtp;
 				struct compat_timespec __user *compat_rmtp;
-#endif
 			};
 			u64 expires;
 		} nanosleep;
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 7ac1bb9ea7da..4df16a70b0d7 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -254,7 +254,7 @@ asmlinkage long sys_adjtimex(struct timex __user *txc_p);
 asmlinkage long sys_times(struct tms __user *tbuf);
 
 asmlinkage long sys_gettid(void);
-asmlinkage long sys_nanosleep(struct timespec __user *rqtp, struct timespec __user *rmtp);
+asmlinkage long sys_nanosleep(struct __kernel_timespec __user *rqtp, struct __kernel_timespec __user *rmtp);
 asmlinkage long sys_alarm(unsigned int seconds);
 asmlinkage long sys_getpid(void);
 asmlinkage long sys_getppid(void);
@@ -317,8 +317,8 @@ asmlinkage long sys_clock_adjtime(clockid_t which_clock,
 asmlinkage long sys_clock_getres(clockid_t which_clock,
 				struct __kernel_timespec __user *tp);
 asmlinkage long sys_clock_nanosleep(clockid_t which_clock, int flags,
-				const struct timespec __user *rqtp,
-				struct timespec __user *rmtp);
+				const struct __kernel_timespec __user *rqtp,
+				struct __kernel_timespec __user *rmtp);
 
 asmlinkage long sys_nice(int increment);
 asmlinkage long sys_sched_setscheduler(pid_t pid, int policy,
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c
index bbd740171554..6cb6e6632ecf 100644
--- a/kernel/time/hrtimer.c
+++ b/kernel/time/hrtimer.c
@@ -1541,8 +1541,10 @@ long hrtimer_nanosleep(const struct timespec64 *rqtp,
 	return ret;
 }
 
-SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp,
-		struct timespec __user *, rmtp)
+#if !defined(CONFIG_64BIT_TIME) || defined(CONFIG_64BIT)
+
+SYSCALL_DEFINE2(nanosleep, struct __kernel_timespec __user *, rqtp,
+		struct __kernel_timespec __user *, rmtp)
 {
 	struct timespec64 tu;
 
@@ -1557,6 +1559,8 @@ SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp,
 	return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC);
 }
 
+#endif
+
 #ifdef CONFIG_COMPAT_32BIT_TIME
 
 COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp,
diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c
index 28c22f54d0f2..f7e7a05e651d 100644
--- a/kernel/time/posix-stubs.c
+++ b/kernel/time/posix-stubs.c
@@ -114,8 +114,8 @@ SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock, struct __kernel_time
 }
 
 SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags,
-		const struct timespec __user *, rqtp,
-		struct timespec __user *, rmtp)
+		const struct __kernel_timespec __user *, rqtp,
+		struct __kernel_timespec __user *, rmtp)
 {
 	struct timespec64 t;
 
diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c
index d46ee8264fd3..a6c4cb6de6a2 100644
--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -1205,8 +1205,8 @@ static int common_nsleep(const clockid_t which_clock, int flags,
 }
 
 SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags,
-		const struct timespec __user *, rqtp,
-		struct timespec __user *, rmtp)
+		const struct __kernel_timespec __user *, rqtp,
+		struct __kernel_timespec __user *, rmtp)
 {
 	const struct k_clock *kc = clockid_to_kclock(which_clock);
 	struct timespec64 t;
-- 
2.14.1

Re: [PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Arnd Bergmann]

On Mon, Nov 27, 2017 at 8:30 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> The series is a preparation series for individual architectures
> to use 64 bit time_t syscalls in compat and 32 bit emulation modes.
>
> This is a follow up to the series Arnd Bergmann posted:
> https://sourceware.org/ml/libc-alpha/2015-05/msg00070.html [1]
>
> Big picture is as per the lwn article:
> https://lwn.net/Articles/643234/ [2]
>
> The series is directed at converting posix clock syscalls:
> clock_gettime, clock_settime, clock_getres and clock_nanosleep
> to use a new data structure __kernel_timespec at syscall boundaries.
> __kernel_timespec maintains 64 bit time_t across all execution modes.
>
> vdso will be handled as part of each architecture when they enable
> support for 64 bit time_t.
>
> The compat syscalls are repurposed to provide backward compatibility
> by using them as native syscalls as well for 32 bit architectures.
> They will continue to use timespec at syscall boundaries.
>
> CONFIG_64_BIT_TIME controls whether the syscalls use __kernel_timespec
> or timespec at syscall boundaries.
>
> The series does the following:
> 1. Enable compat syscalls on 32 bit architectures.
> 2. Add a new __kernel_timespec type to be used as the data structure
>    for all the new syscalls.
> 3. Add new config CONFIG_64BIT_TIME(intead of the CONFIG_COMPAT_TIME in
>    [1] and [2] to switch to new definition of __kernel_timespec. It is
>    the same as struct timespec otherwise.
> 4. Add new CONFIG_32BIT_TIME to conditionally compile compat syscalls.
>
> * Changes since v1:
>  * Introduce CONFIG_32BIT_TIME
>  * Fixed zeroing out of higher order bits of tv_nsec
>  * Included Arnd's changes to fix up use of compat headers

Very nice. I think it would be good to get this into linux-next soon so we
can build on top of this. I have submitted most other y2038 patches today
that don't depend on either this or one of my other patches.

There is one patch that I want to do but haven't imlpemented yet, to merge
get_timespec64() and compat_get_timespec() into one function that
take a bunch of flags (check nanosecond, nano/microsecond,
zero upper half of nanoseconds, 32-bit or 64-bit wide), since I found
a few functions that need more than one of these, and they don't
all need the same combinations. My patch will certainly conflict
with yours, as your touch the same functions, but that's fine.

If you end up doing another version of the series though, it might
be better to move the compat accessors into kernel/time/time.c
along with the native functions, that should make it easier to
consolidate them later.

> I decided against using LEGACY_TIME_SYSCALLS to conditionally compile
> legacy time syscalls such as sys_nanosleep because this will need to
> enclose compat_sys_nanosleep as well. So, defining it as
>
> config LEGACY_TIME_SYSCALLS
>      def_bool 64BIT || !64BIT_TIME
>
> will not include compat_sys_nanosleep. We will instead need a new config to
> exclusively mark legacy syscalls.

Do you mean we would need to do this separately for native and compat
syscalls, and have yet another option, like LEGACY_TIME_SYSCALLS
and LEGACY_TIME_COMPAT_SYSCALLS, to cover all cases? I would
think that CONFIG_COMPAT_32BIT_TIME handles all the compat versions,
while CONFIG_LEGACY_TIME_SYSCALLS handles all the native ones.

       Arnd

Re: [PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Deepa Dinamani]

>> I decided against using LEGACY_TIME_SYSCALLS to conditionally compile
>> legacy time syscalls such as sys_nanosleep because this will need to
>> enclose compat_sys_nanosleep as well. So, defining it as
>>
>> config LEGACY_TIME_SYSCALLS
>>      def_bool 64BIT || !64BIT_TIME
>>
>> will not include compat_sys_nanosleep. We will instead need a new config to
>> exclusively mark legacy syscalls.
>
> Do you mean we would need to do this separately for native and compat
> syscalls, and have yet another option, like LEGACY_TIME_SYSCALLS
> and LEGACY_TIME_COMPAT_SYSCALLS, to cover all cases? I would
> think that CONFIG_COMPAT_32BIT_TIME handles all the compat versions,
> while CONFIG_LEGACY_TIME_SYSCALLS handles all the native ones.

I meant sys_nanosleep would be covered by LEGACY_TIME_SYSCALLS, but
compat_sys_nanosleep would be covered by CONFIG_COMPAT_32BIT_TIME
along with other compat syscalls.
So, if we define the LEGACY_TIME_SYSCALLS as


        "This controls the compilation of the following system calls:
        time, stime, gettimeofday, settimeofday, adjtimex, nanosleep,
alarm, getitimer,
        setitimer, select, utime, utimes, futimesat, and
{old,new}{l,f,}stat{,64}.
        These all pass 32-bit time_t arguments on 32-bit architectures and
        are replaced by other interfaces (e.g. posix timers and clocks, statx).
        C libraries implementing 64-bit time_t in 32-bit architectures have to
        implement the handles by wrapping around the newer interfaces.
        New architectures should not explicitly enable this."

This would not be really true as compat interfaces have nothing to do
with this config.

I was proposing that we could have LEGACY_TIME_SYSCALLS config, but
then have all these "deprecated" syscalls be enclosed within this,
compat or not.
This will also mean that we will have to come up representing these
syscalls in the syscall header files.
This can be a separate patch and this series can be merged as is if
everyone agrees.

-Deepa

Re: [PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Arnd Bergmann]

On Mon, Nov 27, 2017 at 11:29 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
>>> I decided against using LEGACY_TIME_SYSCALLS to conditionally compile
>>> legacy time syscalls such as sys_nanosleep because this will need to
>>> enclose compat_sys_nanosleep as well. So, defining it as
>>>
>>> config LEGACY_TIME_SYSCALLS
>>>      def_bool 64BIT || !64BIT_TIME
>>>
>>> will not include compat_sys_nanosleep. We will instead need a new config to
>>> exclusively mark legacy syscalls.
>>
>> Do you mean we would need to do this separately for native and compat
>> syscalls, and have yet another option, like LEGACY_TIME_SYSCALLS
>> and LEGACY_TIME_COMPAT_SYSCALLS, to cover all cases? I would
>> think that CONFIG_COMPAT_32BIT_TIME handles all the compat versions,
>> while CONFIG_LEGACY_TIME_SYSCALLS handles all the native ones.
>
> I meant sys_nanosleep would be covered by LEGACY_TIME_SYSCALLS, but
> compat_sys_nanosleep would be covered by CONFIG_COMPAT_32BIT_TIME
> along with other compat syscalls.
> So, if we define the LEGACY_TIME_SYSCALLS as
>
>
>         "This controls the compilation of the following system calls:
>         time, stime, gettimeofday, settimeofday, adjtimex, nanosleep,
> alarm, getitimer,
>         setitimer, select, utime, utimes, futimesat, and
> {old,new}{l,f,}stat{,64}.
>         These all pass 32-bit time_t arguments on 32-bit architectures and
>         are replaced by other interfaces (e.g. posix timers and clocks, statx).
>         C libraries implementing 64-bit time_t in 32-bit architectures have to
>         implement the handles by wrapping around the newer interfaces.
>         New architectures should not explicitly enable this."
>
> This would not be really true as compat interfaces have nothing to do
> with this config.
>
> I was proposing that we could have LEGACY_TIME_SYSCALLS config, but
> then have all these "deprecated" syscalls be enclosed within this,
> compat or not.
> This will also mean that we will have to come up representing these
> syscalls in the syscall header files.
> This can be a separate patch and this series can be merged as is if
> everyone agrees.

I think doing this separately  would be good, I don't see any interdependency
with the other patches, we just need to decide what we want in the long
run.

I agree my text that you cited doesn't capture the situation correctly,
as this is really about the obsolete system calls that take 64-bit time_t
arguments on architectures that are converted to allow 64-bit time_t
for non-obsolete system calls.

Maybe it's better to just reword this to

      "This controls the compilation of the following system calls:
      time, stime, gettimeofday, settimeofday, adjtimex, nanosleep,
alarm, getitimer,
      setitimer, select, utime, utimes, futimesat, and {old,new}{l,f,}stat{,64}.
      These are all replaced by other interfaces (e.g. posix timers and clocks,
      statx) on architectures that got converted from 32-bit time_t to
64-bit time_t.
      C libraries implementing 64-bit time_t in 32-bit architectures have to
      implement the handles by wrapping around the newer interfaces.
      New architectures should not explicitly enable this."

That would clarify that it's not about the compat system calls, while
also allowing the two options to be set independently.

        Arnd

Re: [PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Deepa Dinamani]

On Tue, Nov 28, 2017 at 6:17 AM, Arnd Bergmann <arnd@arndb.de> wrote:
> On Mon, Nov 27, 2017 at 11:29 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
>>>> I decided against using LEGACY_TIME_SYSCALLS to conditionally compile
>>>> legacy time syscalls such as sys_nanosleep because this will need to
>>>> enclose compat_sys_nanosleep as well. So, defining it as
>>>>
>>>> config LEGACY_TIME_SYSCALLS
>>>>      def_bool 64BIT || !64BIT_TIME
>>>>
>>>> will not include compat_sys_nanosleep. We will instead need a new config to
>>>> exclusively mark legacy syscalls.
>>>
>>> Do you mean we would need to do this separately for native and compat
>>> syscalls, and have yet another option, like LEGACY_TIME_SYSCALLS
>>> and LEGACY_TIME_COMPAT_SYSCALLS, to cover all cases? I would
>>> think that CONFIG_COMPAT_32BIT_TIME handles all the compat versions,
>>> while CONFIG_LEGACY_TIME_SYSCALLS handles all the native ones.
>>
>> I meant sys_nanosleep would be covered by LEGACY_TIME_SYSCALLS, but
>> compat_sys_nanosleep would be covered by CONFIG_COMPAT_32BIT_TIME
>> along with other compat syscalls.
>> So, if we define the LEGACY_TIME_SYSCALLS as
>>
>>
>>         "This controls the compilation of the following system calls:
>>         time, stime, gettimeofday, settimeofday, adjtimex, nanosleep,
>> alarm, getitimer,
>>         setitimer, select, utime, utimes, futimesat, and
>> {old,new}{l,f,}stat{,64}.
>>         These all pass 32-bit time_t arguments on 32-bit architectures and
>>         are replaced by other interfaces (e.g. posix timers and clocks, statx).
>>         C libraries implementing 64-bit time_t in 32-bit architectures have to
>>         implement the handles by wrapping around the newer interfaces.
>>         New architectures should not explicitly enable this."
>>
>> This would not be really true as compat interfaces have nothing to do
>> with this config.
>>
>> I was proposing that we could have LEGACY_TIME_SYSCALLS config, but
>> then have all these "deprecated" syscalls be enclosed within this,
>> compat or not.
>> This will also mean that we will have to come up representing these
>> syscalls in the syscall header files.
>> This can be a separate patch and this series can be merged as is if
>> everyone agrees.
>
> I think doing this separately  would be good, I don't see any interdependency
> with the other patches, we just need to decide what we want in the long
> run.

Right. There are three options:

1. Use two configs to identify which syscalls need not be supported by
new architectures.
In this case it makes sense to say LEGACY_TIME_SYSCALLS and
COMPAT_32BIT_TIME both need to be disabled for new architectures. And,
I can reword the config to what you mention below.

2. Make the LEGACY_TIME_SYSCALLS eliminate non y2038 safe syscalls
mentioned below only.
In this case only the native and compat functions of the below
mentioned syscalls need to be identified by the config. I like this
option as this clearly identifies which syscalls are deprecated and do
not have a 64 bit counterpart. Not all architectures need to support
turning this off.

3. If we don't need either 1 or 2, then we could stick with what we
have today in the series as CONFIG_64BIT_TIME will be deleted and they
only need #ifdef CONFIG_64BIT.

Let me know if anyone prefers something else.

> I agree my text that you cited doesn't capture the situation correctly,
> as this is really about the obsolete system calls that take 64-bit time_t
> arguments on architectures that are converted to allow 64-bit time_t
> for non-obsolete system calls.
>
> Maybe it's better to just reword this to
>
>       "This controls the compilation of the following system calls:
>       time, stime, gettimeofday, settimeofday, adjtimex, nanosleep,
> alarm, getitimer,
>       setitimer, select, utime, utimes, futimesat, and {old,new}{l,f,}stat{,64}.
>       These are all replaced by other interfaces (e.g. posix timers and clocks,
>       statx) on architectures that got converted from 32-bit time_t to
> 64-bit time_t.
>       C libraries implementing 64-bit time_t in 32-bit architectures have to
>       implement the handles by wrapping around the newer interfaces.
>       New architectures should not explicitly enable this."
>
> That would clarify that it's not about the compat system calls, while
> also allowing the two options to be set independently.

-Deepa

Re: [PATCH v2 00/10] posix_clocks: Prepare syscalls for 64 bit time_t conversion

[Arnd Bergmann]

On Wed, Nov 29, 2017 at 12:17 AM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> On Tue, Nov 28, 2017 at 6:17 AM, Arnd Bergmann <arnd@arndb.de> wrote:
>> On Mon, Nov 27, 2017 at 11:29 PM, Deepa Dinamani <deepa.kernel@gmail.com> wrote:
> Right. There are three options:
>
> 1. Use two configs to identify which syscalls need not be supported by
> new architectures.
> In this case it makes sense to say LEGACY_TIME_SYSCALLS and
> COMPAT_32BIT_TIME both need to be disabled for new architectures. And,
> I can reword the config to what you mention below.
>
> 2. Make the LEGACY_TIME_SYSCALLS eliminate non y2038 safe syscalls
> mentioned below only.
> In this case only the native and compat functions of the below
> mentioned syscalls need to be identified by the config. I like this
> option as this clearly identifies which syscalls are deprecated and do
> not have a 64 bit counterpart. Not all architectures need to support
> turning this off.
>
> 3. If we don't need either 1 or 2, then we could stick with what we
> have today in the series as CONFIG_64BIT_TIME will be deleted and they
> only need #ifdef CONFIG_64BIT.
>
> Let me know if anyone prefers something else.

I think I prefer to have both LEGACY_TIME_SYSCALLS to guard
the native deprecated syscalls (disabled on 32-bit architectures after
the conversion, and enabled on 64-bit architectures until
we merge the next one), and COMPAT_32BIT_TIME to guard the
compat versions of both the deprecated and the non-deprecated
syscalls (enabled on all existing 32-bit architectures after the
conversion, and on 64-bit architectures if they provide a compat
mode for the former).

Those two are not symmetric, but I think those are the most
common combinations, and the Kconfig symbol helps document
what they are.

There is one more category for things like io_getevents() and
rt_sigtimedwait that also need two separate compat versions,
one for 32-bit time_t and one for 64-bit time_t, but it seems better
to deal with those case-by-case rather than introducing another
Kconfig symbol.

        Arnd

Re: [Y2038] [PATCH v2 10/10] nanosleep: change time types to safe __kernel_* types

[Ben Hutchings]

On Mon, 2017-11-27 at 11:30 -0800, Deepa Dinamani wrote:
[...]
> @@ -15,9 +16,7 @@ struct pollfd;
>  enum timespec_type {
>  	TT_NONE		= 0,
>  	TT_NATIVE	= 1,
> -#ifdef CONFIG_COMPAT
>  	TT_COMPAT	= 2,
> -#endif
>  };
>  
>  /*
> @@ -40,10 +39,8 @@ struct restart_block {
>  			clockid_t clockid;
>  			enum timespec_type type;
>  			union {
> -				struct timespec __user *rmtp;
> -#ifdef CONFIG_COMPAT
> +				struct __kernel_timespec __user *rmtp;
>  				struct compat_timespec __user *compat_rmtp;
> -#endif
>  			};
>  			u64 expires;
>  		} nanosleep;
[...]

The deletions of #ifdef and #endif here seem to belong in patch #6, as
that's the one that changes the condition for the compat nanosleep
implementation which uses the compat_rmtp field.

Ben.

-- 
Ben Hutchings
Software Developer, Codethink Ltd.