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
Big picture is as per the lwn article: https://lwn.net/Articles/643234/
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 unconditionally. 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.
Arnd Bergmann (1): y2038: introduce CONFIG_64BIT_TIME
Deepa Dinamani (8): include: Move compat_timespec/ timeval to compat_time.h compat: Make compat helpers independent of CONFIG_COMPAT compat: enable compat_get/put_timespec64 always posix-clocks: Enable compat syscalls always 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 | 11 ++++ 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 | 7 ++- 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 | 7 +-- kernel/time/posix-stubs.c | 12 ++--- kernel/time/posix-timers.c | 20 ++++---- kernel/time/time.c | 10 +++- 45 files changed, 152 insertions(+), 195 deletions(-) create mode 100644 include/linux/compat_time.h
base-commit: d9e0e63d9a6f88440eb201e1491fcf730272c706
All the current architecture specific defines for these are the same. Refactor these common defines to a common header file.
The new common linux/compat_time.h is also useful as it will eventually be used to hold all the defines that are needed for compat time types that support non y2038 safe types. New architectures need not have to define these new types as they will only use new y2038 safe syscalls. This file can be deleted after y2038 when we stop supporting non y2038 safe syscalls.
The patch also requires an operation similar to:
git grep "asm/compat.h" | cut -d ":" -f 1 | xargs -n 1 sed -i -e "s%asm/compat.h%linux/compat.h%g"
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-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 Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- 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 | 1 + include/linux/compat_time.h | 19 +++++++++++++++++++ 32 files changed, 32 insertions(+), 110 deletions(-) create mode 100644 include/linux/compat_time.h
diff --git a/arch/arm64/include/asm/compat.h b/arch/arm64/include/asm/compat.h index e39d487bf724..d4f9c9ee3b15 100644 --- a/arch/arm64/include/asm/compat.h +++ b/arch/arm64/include/asm/compat.h @@ -34,7 +34,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u16 __compat_uid_t; @@ -66,16 +65,6 @@ typedef u32 compat_ulong_t; typedef u64 compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { #ifdef __AARCH64EB__ short st_dev; diff --git a/arch/arm64/include/asm/stat.h b/arch/arm64/include/asm/stat.h index 15e35598ac40..eab738019707 100644 --- a/arch/arm64/include/asm/stat.h +++ b/arch/arm64/include/asm/stat.h @@ -20,6 +20,7 @@
#ifdef CONFIG_COMPAT
+#include <linux/compat_time.h> #include <asm/compat.h>
/* diff --git a/arch/arm64/kernel/hw_breakpoint.c b/arch/arm64/kernel/hw_breakpoint.c index 749f81779420..bfa2b78cf0e3 100644 --- a/arch/arm64/kernel/hw_breakpoint.c +++ b/arch/arm64/kernel/hw_breakpoint.c @@ -29,7 +29,6 @@ #include <linux/ptrace.h> #include <linux/smp.h>
-#include <asm/compat.h> #include <asm/current.h> #include <asm/debug-monitors.h> #include <asm/hw_breakpoint.h> diff --git a/arch/arm64/kernel/perf_regs.c b/arch/arm64/kernel/perf_regs.c index 1d091d048d04..929fc369d0be 100644 --- a/arch/arm64/kernel/perf_regs.c +++ b/arch/arm64/kernel/perf_regs.c @@ -5,7 +5,7 @@ #include <linux/bug.h> #include <linux/sched/task_stack.h>
-#include <asm/compat.h> +#include <linux/compat.h> #include <asm/perf_regs.h> #include <asm/ptrace.h>
diff --git a/arch/arm64/kernel/process.c b/arch/arm64/kernel/process.c index b2adcce7bc18..1acb3097d35b 100644 --- a/arch/arm64/kernel/process.c +++ b/arch/arm64/kernel/process.c @@ -52,7 +52,6 @@ #include <linux/thread_info.h>
#include <asm/alternative.h> -#include <asm/compat.h> #include <asm/cacheflush.h> #include <asm/exec.h> #include <asm/fpsimd.h> diff --git a/arch/mips/include/asm/compat.h b/arch/mips/include/asm/compat.h index 8e2b5b556488..ebbf3f04f82b 100644 --- a/arch/mips/include/asm/compat.h +++ b/arch/mips/include/asm/compat.h @@ -14,7 +14,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_suseconds_t;
@@ -46,16 +45,6 @@ typedef u32 compat_ulong_t; typedef u64 compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; s32 st_pad1[3]; diff --git a/arch/mips/kernel/signal32.c b/arch/mips/kernel/signal32.c index cf5c7c05e5a3..a6b04c70a8cb 100644 --- a/arch/mips/kernel/signal32.c +++ b/arch/mips/kernel/signal32.c @@ -14,7 +14,7 @@ #include <linux/signal.h> #include <linux/syscalls.h>
-#include <asm/compat.h> +#include <linux/compat.h> #include <asm/compat-signal.h> #include <linux/uaccess.h> #include <asm/unistd.h> diff --git a/arch/parisc/include/asm/compat.h b/arch/parisc/include/asm/compat.h index 07f48827afda..ea453c356bf9 100644 --- a/arch/parisc/include/asm/compat.h +++ b/arch/parisc/include/asm/compat.h @@ -13,7 +13,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u32 __compat_uid_t; @@ -40,16 +39,6 @@ typedef u32 compat_ulong_t; typedef u64 compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; /* dev_t is 32 bits on parisc */ compat_ino_t st_ino; /* 32 bits */ diff --git a/arch/powerpc/include/asm/compat.h b/arch/powerpc/include/asm/compat.h index a035b1e5dfa7..d48892bcd38f 100644 --- a/arch/powerpc/include/asm/compat.h +++ b/arch/powerpc/include/asm/compat.h @@ -17,7 +17,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u32 __compat_uid_t; @@ -45,16 +44,6 @@ typedef u32 compat_ulong_t; typedef u64 compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; compat_ino_t st_ino; diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 4ce0e8eccbcf..7698bd054181 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -42,7 +42,7 @@ #include <asm/paca.h> #include <asm/lppaca.h> #include <asm/cache.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/mmu.h> #include <asm/hvcall.h> #include <asm/xics.h> diff --git a/arch/powerpc/oprofile/backtrace.c b/arch/powerpc/oprofile/backtrace.c index ecc66d5f02c9..11ff763c03ad 100644 --- a/arch/powerpc/oprofile/backtrace.c +++ b/arch/powerpc/oprofile/backtrace.c @@ -11,7 +11,7 @@ #include <linux/sched.h> #include <asm/processor.h> #include <linux/uaccess.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/oprofile_impl.h>
#define STACK_SP(STACK) *(STACK) diff --git a/arch/s390/hypfs/hypfs_sprp.c b/arch/s390/hypfs/hypfs_sprp.c index ae0ed8dd5f1b..5d85a039391c 100644 --- a/arch/s390/hypfs/hypfs_sprp.c +++ b/arch/s390/hypfs/hypfs_sprp.c @@ -13,7 +13,6 @@ #include <linux/string.h> #include <linux/types.h> #include <linux/uaccess.h> -#include <asm/compat.h> #include <asm/diag.h> #include <asm/sclp.h> #include "hypfs.h" diff --git a/arch/s390/include/asm/compat.h b/arch/s390/include/asm/compat.h index 1b60eb3676d5..32a77b6d9a59 100644 --- a/arch/s390/include/asm/compat.h +++ b/arch/s390/include/asm/compat.h @@ -53,7 +53,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u16 __compat_uid_t; @@ -97,16 +96,6 @@ typedef struct { u32 gprs_high[NUM_GPRS]; } s390_compat_regs_high;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; u16 __pad1; diff --git a/arch/s390/include/asm/elf.h b/arch/s390/include/asm/elf.h index 9a3cb3983c01..78f75384f891 100644 --- a/arch/s390/include/asm/elf.h +++ b/arch/s390/include/asm/elf.h @@ -126,7 +126,7 @@ */
#include <asm/ptrace.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/syscall.h> #include <asm/user.h>
@@ -136,7 +136,6 @@ typedef s390_regs elf_gregset_t; typedef s390_fp_regs compat_elf_fpregset_t; typedef s390_compat_regs compat_elf_gregset_t;
-#include <linux/compat.h> #include <linux/sched/mm.h> /* for task_struct */ #include <asm/mmu_context.h>
diff --git a/arch/s390/kvm/priv.c b/arch/s390/kvm/priv.c index c954ac49eee4..07eebba4bd37 100644 --- a/arch/s390/kvm/priv.c +++ b/arch/s390/kvm/priv.c @@ -29,7 +29,6 @@ #include <asm/gmap.h> #include <asm/io.h> #include <asm/ptrace.h> -#include <asm/compat.h> #include <asm/sclp.h> #include "gaccess.h" #include "kvm-s390.h" diff --git a/arch/s390/pci/pci_clp.c b/arch/s390/pci/pci_clp.c index 93cd0f1ca12b..19b2d2a9b43d 100644 --- a/arch/s390/pci/pci_clp.c +++ b/arch/s390/pci/pci_clp.c @@ -19,7 +19,6 @@ #include <linux/uaccess.h> #include <asm/pci_debug.h> #include <asm/pci_clp.h> -#include <asm/compat.h> #include <asm/clp.h> #include <uapi/asm/clp.h>
diff --git a/arch/sparc/include/asm/compat.h b/arch/sparc/include/asm/compat.h index 977c3f280ba1..7348f111d169 100644 --- a/arch/sparc/include/asm/compat.h +++ b/arch/sparc/include/asm/compat.h @@ -11,7 +11,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u16 __compat_uid_t; @@ -39,16 +38,6 @@ typedef u32 compat_ulong_t; typedef u64 compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; compat_ino_t st_ino; diff --git a/arch/tile/include/asm/compat.h b/arch/tile/include/asm/compat.h index c14e36f008c8..c3a326c9ae75 100644 --- a/arch/tile/include/asm/compat.h +++ b/arch/tile/include/asm/compat.h @@ -29,7 +29,6 @@ typedef u32 compat_ulong_t; typedef u32 compat_size_t; typedef s32 compat_ssize_t; typedef s32 compat_off_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef u32 compat_ino_t; typedef u32 compat_caddr_t; @@ -59,16 +58,6 @@ typedef unsigned long compat_elf_greg_t; #define COMPAT_ELF_NGREG (sizeof(struct pt_regs) / sizeof(compat_elf_greg_t)) typedef compat_elf_greg_t compat_elf_gregset_t[COMPAT_ELF_NGREG];
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - #define compat_stat stat #define compat_statfs statfs
diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c index 140d33288e78..6b8961912781 100644 --- a/arch/x86/events/core.c +++ b/arch/x86/events/core.c @@ -2391,7 +2391,7 @@ static unsigned long get_segment_base(unsigned int segment)
#ifdef CONFIG_IA32_EMULATION
-#include <asm/compat.h> +#include <linux/compat.h>
static inline int perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry_ctx *entry) diff --git a/arch/x86/include/asm/compat.h b/arch/x86/include/asm/compat.h index a600a6cda9ec..5d8e0fc204db 100644 --- a/arch/x86/include/asm/compat.h +++ b/arch/x86/include/asm/compat.h @@ -17,7 +17,6 @@
typedef u32 compat_size_t; typedef s32 compat_ssize_t; -typedef s32 compat_time_t; typedef s32 compat_clock_t; typedef s32 compat_pid_t; typedef u16 __compat_uid_t; @@ -46,16 +45,6 @@ typedef u32 compat_u32; typedef u64 __attribute__((aligned(4))) compat_u64; typedef u32 compat_uptr_t;
-struct compat_timespec { - compat_time_t tv_sec; - s32 tv_nsec; -}; - -struct compat_timeval { - compat_time_t tv_sec; - s32 tv_usec; -}; - struct compat_stat { compat_dev_t st_dev; u16 __pad1; diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 09ad88572746..db25aa15b705 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -49,7 +49,7 @@ int ftrace_int3_handler(struct pt_regs *regs); #if !defined(__ASSEMBLY__) && !defined(COMPILE_OFFSETS)
#if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_IA32_EMULATION) -#include <asm/compat.h> +#include <linux/compat.h>
/* * Because ia32 syscalls do not map to x86_64 syscall numbers diff --git a/arch/x86/include/asm/sys_ia32.h b/arch/x86/include/asm/sys_ia32.h index 82c34ee25a65..8527b26ad36f 100644 --- a/arch/x86/include/asm/sys_ia32.h +++ b/arch/x86/include/asm/sys_ia32.h @@ -16,7 +16,7 @@ #include <linux/linkage.h> #include <linux/types.h> #include <linux/signal.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/ia32.h>
/* ia32/sys_ia32.c */ diff --git a/arch/x86/kernel/sys_x86_64.c b/arch/x86/kernel/sys_x86_64.c index a63fe77b3217..13d51bd80409 100644 --- a/arch/x86/kernel/sys_x86_64.c +++ b/arch/x86/kernel/sys_x86_64.c @@ -19,7 +19,7 @@ #include <linux/elf.h>
#include <asm/elf.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/ia32.h> #include <asm/syscalls.h> #include <asm/mpx.h> diff --git a/drivers/s390/block/dasd_ioctl.c b/drivers/s390/block/dasd_ioctl.c index 7bdc6aaa0ba3..2016e0ed5865 100644 --- a/drivers/s390/block/dasd_ioctl.c +++ b/drivers/s390/block/dasd_ioctl.c @@ -18,7 +18,6 @@ #include <linux/fs.h> #include <linux/blkpg.h> #include <linux/slab.h> -#include <asm/compat.h> #include <asm/ccwdev.h> #include <asm/schid.h> #include <asm/cmb.h> diff --git a/drivers/s390/char/fs3270.c b/drivers/s390/char/fs3270.c index c4518168fd02..d74facaa1755 100644 --- a/drivers/s390/char/fs3270.c +++ b/drivers/s390/char/fs3270.c @@ -18,7 +18,6 @@ #include <linux/slab.h> #include <linux/types.h>
-#include <asm/compat.h> #include <asm/ccwdev.h> #include <asm/cio.h> #include <asm/ebcdic.h> diff --git a/drivers/s390/char/sclp_ctl.c b/drivers/s390/char/sclp_ctl.c index a78cea0c3a09..248b5db3eaa8 100644 --- a/drivers/s390/char/sclp_ctl.c +++ b/drivers/s390/char/sclp_ctl.c @@ -14,7 +14,6 @@ #include <linux/init.h> #include <linux/ioctl.h> #include <linux/fs.h> -#include <asm/compat.h> #include <asm/sclp_ctl.h> #include <asm/sclp.h>
diff --git a/drivers/s390/char/vmcp.c b/drivers/s390/char/vmcp.c index 17e411c57576..948ce82a7725 100644 --- a/drivers/s390/char/vmcp.c +++ b/drivers/s390/char/vmcp.c @@ -23,7 +23,6 @@ #include <linux/mutex.h> #include <linux/cma.h> #include <linux/mm.h> -#include <asm/compat.h> #include <asm/cpcmd.h> #include <asm/debug.h> #include <asm/vmcp.h> diff --git a/drivers/s390/cio/chsc_sch.c b/drivers/s390/cio/chsc_sch.c index 8e7e19b9e92c..662d24c6b79c 100644 --- a/drivers/s390/cio/chsc_sch.c +++ b/drivers/s390/cio/chsc_sch.c @@ -15,7 +15,6 @@ #include <linux/miscdevice.h> #include <linux/kernel_stat.h>
-#include <asm/compat.h> #include <asm/cio.h> #include <asm/chsc.h> #include <asm/isc.h> diff --git a/drivers/s390/net/qeth_core_main.c b/drivers/s390/net/qeth_core_main.c index 49b9efeba1bd..8259c0ae3395 100644 --- a/drivers/s390/net/qeth_core_main.c +++ b/drivers/s390/net/qeth_core_main.c @@ -26,7 +26,7 @@ #include <asm/chpid.h> #include <asm/io.h> #include <asm/sysinfo.h> -#include <asm/compat.h> +#include <linux/compat.h> #include <asm/diag.h> #include <asm/cio.h> #include <asm/ccwdev.h> diff --git a/drivers/staging/pi433/pi433_if.c b/drivers/staging/pi433/pi433_if.c index a960fe2e7875..3fcb238b1995 100644 --- a/drivers/staging/pi433/pi433_if.c +++ b/drivers/staging/pi433/pi433_if.c @@ -48,7 +48,7 @@ #include <linux/wait.h> #include <linux/spi/spi.h> #ifdef CONFIG_COMPAT -#include <asm/compat.h> +#include <linux/compat.h> #endif
#include "pi433_if.h" diff --git a/include/linux/compat.h b/include/linux/compat.h index 0fc36406f32c..1a5e397ffcd4 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -7,6 +7,7 @@ */
#include <linux/types.h> +#include <linux/compat_time.h>
#ifdef CONFIG_COMPAT
diff --git a/include/linux/compat_time.h b/include/linux/compat_time.h new file mode 100644 index 000000000000..56a54a1e4355 --- /dev/null +++ b/include/linux/compat_time.h @@ -0,0 +1,19 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_COMPAT_TIME_H +#define _LINUX_COMPAT_TIME_H + +#include <linux/types.h> + +typedef s32 compat_time_t; + +struct compat_timespec { + compat_time_t tv_sec; + s32 tv_nsec; +}; + +struct compat_timeval { + compat_time_t tv_sec; + s32 tv_usec; +}; + +#endif /* _LINUX_COMPAT_TIME_H */
On Fri, 10 Nov 2017 14:42:51 -0800 Deepa Dinamani deepa.kernel@gmail.com wrote:
diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index 09ad88572746..db25aa15b705 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -49,7 +49,7 @@ int ftrace_int3_handler(struct pt_regs *regs); #if !defined(__ASSEMBLY__) && !defined(COMPILE_OFFSETS) #if defined(CONFIG_FTRACE_SYSCALLS) && defined(CONFIG_IA32_EMULATION) -#include <asm/compat.h> +#include <linux/compat.h>
Acked-by: Steven Rostedt (VMware) rostedt@goodmis.org
-- Steve
Many of the compat time syscalls are also repurposed as 32 bit native syscalls to provide backward compatibility while adding new y2038 safe sycalls. Enabling the helpers makes this possible.
Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- include/linux/compat.h | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/include/linux/compat.h b/include/linux/compat.h index 1a5e397ffcd4..f59f1f2d03de 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -24,6 +24,8 @@ #include <asm/siginfo.h> #include <asm/signal.h>
+#endif + #ifndef COMPAT_USE_64BIT_TIME #define COMPAT_USE_64BIT_TIME 0 #endif @@ -59,6 +61,8 @@ } \ static inline long C_SYSC##name(__MAP(x,__SC_DECL,__VA_ARGS__))
+#ifdef CONFIG_COMPAT + #ifndef compat_user_stack_pointer #define compat_user_stack_pointer() current_user_stack_pointer() #endif
On Fri, Nov 10, 2017 at 11:42 PM, Deepa Dinamani deepa.kernel@gmail.com wrote:
Many of the compat time syscalls are also repurposed as 32 bit native syscalls to provide backward compatibility while adding new y2038 safe sycalls. Enabling the helpers makes this possible.
Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com
I ran into a build error on ARM64 with CONFIG_COMPAT=n with this patch, here is a fixup that makes it build cleanly and should also work on other architectures. Unfortunately, including asm/compat.h has a number of dependencies on other headers, so I decided to just include them all for the fixup.
Alternatively, we could change the architecture code to always include asm/compat.h when using is_compat_thread(), or to replace is_compat_thread() calls with in_compat_syscall(), but the approach below should be more likely to work on the 64-bit architectures that I did not try.
Arnd
--- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -9,8 +9,6 @@ #include <linux/types.h> #include <linux/compat_time.h>
-#ifdef CONFIG_COMPAT - #include <linux/stat.h> #include <linux/param.h> /* for HZ */ #include <linux/sem.h> @@ -20,12 +18,12 @@ #include <linux/aio_abi.h> /* for aio_context_t */ #include <linux/unistd.h>
+#ifdef CONFIG_64BIT #include <asm/compat.h> +#endif #include <asm/siginfo.h> #include <asm/signal.h>
-#endif - #ifndef COMPAT_USE_64BIT_TIME #define COMPAT_USE_64BIT_TIME 0 #endif @@ -774,7 +772,9 @@ static inline struct compat_timeval ns_to_compat_timeval(s64 nsec) #else /* !CONFIG_COMPAT */
#define is_compat_task() (0) +#ifndef in_compat_syscall static inline bool in_compat_syscall(void) { return false; } +#endif
#endif /* CONFIG_COMPAT */
Thanks, I will include the patch in the update.
These functions are used in the repurposed compat syscalls to provide backward compatibility for using 32 bit time_t on 32 bit systems.
Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- include/linux/compat.h | 2 - include/linux/compat_time.h | 4 ++ kernel/Makefile | 2 +- kernel/compat.c | 92 +++++++++++++++++++++++---------------------- 4 files changed, 53 insertions(+), 47 deletions(-)
diff --git a/include/linux/compat.h b/include/linux/compat.h index f59f1f2d03de..d7f566c14e84 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -170,8 +170,6 @@ extern int compat_get_timespec(struct timespec *, const void __user *); extern int compat_put_timespec(const struct timespec *, void __user *); extern int compat_get_timeval(struct timeval *, const void __user *); extern int compat_put_timeval(const struct timeval *, void __user *); -extern int compat_get_timespec64(struct timespec64 *, const void __user *); -extern int compat_put_timespec64(const struct timespec64 *, void __user *); extern int get_compat_itimerspec64(struct itimerspec64 *its, const struct compat_itimerspec __user *uits); extern int put_compat_itimerspec64(const struct itimerspec64 *its, diff --git a/include/linux/compat_time.h b/include/linux/compat_time.h index 56a54a1e4355..31f2774f1994 100644 --- a/include/linux/compat_time.h +++ b/include/linux/compat_time.h @@ -3,6 +3,7 @@ #define _LINUX_COMPAT_TIME_H
#include <linux/types.h> +#include <linux/time64.h>
typedef s32 compat_time_t;
@@ -16,4 +17,7 @@ struct compat_timeval { s32 tv_usec; };
+extern int compat_get_timespec64(struct timespec64 *, const void __user *); +extern int compat_put_timespec64(const struct timespec64 *, void __user *); + #endif /* _LINUX_COMPAT_TIME_H */ diff --git a/kernel/Makefile b/kernel/Makefile index 172d151d429c..2e3df3deee18 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -66,7 +66,7 @@ obj-$(CONFIG_KEXEC_CORE) += kexec_core.o obj-$(CONFIG_KEXEC) += kexec.o obj-$(CONFIG_KEXEC_FILE) += kexec_file.o obj-$(CONFIG_BACKTRACE_SELF_TEST) += backtracetest.o -obj-$(CONFIG_COMPAT) += compat.o +obj-y += compat.o obj-$(CONFIG_CGROUPS) += cgroup/ obj-$(CONFIG_UTS_NS) += utsname.o obj-$(CONFIG_USER_NS) += user_namespace.o diff --git a/kernel/compat.c b/kernel/compat.c index d1cee656a7ed..42653ac977c0 100644 --- a/kernel/compat.c +++ b/kernel/compat.c @@ -30,6 +30,52 @@
#include <linux/uaccess.h>
+static int __compat_get_timespec64(struct timespec64 *ts64, + const struct compat_timespec __user *cts) +{ + struct compat_timespec ts; + int ret; + + ret = copy_from_user(&ts, cts, sizeof(ts)); + if (ret) + return -EFAULT; + + ts64->tv_sec = ts.tv_sec; + ts64->tv_nsec = ts.tv_nsec; + + return 0; +} + +static int __compat_put_timespec64(const struct timespec64 *ts64, + struct compat_timespec __user *cts) +{ + struct compat_timespec ts = { + .tv_sec = ts64->tv_sec, + .tv_nsec = ts64->tv_nsec + }; + return copy_to_user(cts, &ts, sizeof(ts)) ? -EFAULT : 0; +} + +int compat_get_timespec64(struct timespec64 *ts, const void __user *uts) +{ + if (COMPAT_USE_64BIT_TIME) + return copy_from_user(ts, uts, sizeof(*ts)) ? -EFAULT : 0; + else + return __compat_get_timespec64(ts, uts); +} +EXPORT_SYMBOL_GPL(compat_get_timespec64); + +int compat_put_timespec64(const struct timespec64 *ts, void __user *uts) +{ + if (COMPAT_USE_64BIT_TIME) + return copy_to_user(uts, ts, sizeof(*ts)) ? -EFAULT : 0; + else + return __compat_put_timespec64(ts, uts); +} +EXPORT_SYMBOL_GPL(compat_put_timespec64); + +#ifdef CONFIG_COMPAT + int compat_get_timex(struct timex *txc, const struct compat_timex __user *utp) { struct compat_timex tx32; @@ -120,50 +166,6 @@ static int __compat_put_timespec(const struct timespec *ts, struct compat_timesp __put_user(ts->tv_nsec, &cts->tv_nsec)) ? -EFAULT : 0; }
-static int __compat_get_timespec64(struct timespec64 *ts64, - const struct compat_timespec __user *cts) -{ - struct compat_timespec ts; - int ret; - - ret = copy_from_user(&ts, cts, sizeof(ts)); - if (ret) - return -EFAULT; - - ts64->tv_sec = ts.tv_sec; - ts64->tv_nsec = ts.tv_nsec; - - return 0; -} - -static int __compat_put_timespec64(const struct timespec64 *ts64, - struct compat_timespec __user *cts) -{ - struct compat_timespec ts = { - .tv_sec = ts64->tv_sec, - .tv_nsec = ts64->tv_nsec - }; - return copy_to_user(cts, &ts, sizeof(ts)) ? -EFAULT : 0; -} - -int compat_get_timespec64(struct timespec64 *ts, const void __user *uts) -{ - if (COMPAT_USE_64BIT_TIME) - return copy_from_user(ts, uts, sizeof(*ts)) ? -EFAULT : 0; - else - return __compat_get_timespec64(ts, uts); -} -EXPORT_SYMBOL_GPL(compat_get_timespec64); - -int compat_put_timespec64(const struct timespec64 *ts, void __user *uts) -{ - if (COMPAT_USE_64BIT_TIME) - return copy_to_user(uts, ts, sizeof(*ts)) ? -EFAULT : 0; - else - return __compat_put_timespec64(ts, uts); -} -EXPORT_SYMBOL_GPL(compat_put_timespec64); - int compat_get_timeval(struct timeval *tv, const void __user *utv) { if (COMPAT_USE_64BIT_TIME) @@ -582,3 +584,5 @@ void __user *compat_alloc_user_space(unsigned long len) return ptr; } EXPORT_SYMBOL_GPL(compat_alloc_user_space); + +#endif
clock_gettime, clock_settime, clock_getres and clock_nanosleep compat syscalls are also repurposed to provide backward compatibility to support 32 bit time_t on 32 bit systems.
Note that nanosleep compat syscall will also be treated the same way as the above syscalls as it shares common handler funtions with clock_nanosleep. But, there is no plan to provide y2038 safe solution for nanosleep.
Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- kernel/time/hrtimer.c | 3 --- kernel/time/posix-stubs.c | 2 +- kernel/time/posix-timers.c | 10 ++++------ 3 files changed, 5 insertions(+), 10 deletions(-)
diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c index d32520840fde..e7db66af5a23 100644 --- a/kernel/time/hrtimer.c +++ b/kernel/time/hrtimer.c @@ -1557,8 +1557,6 @@ SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp, return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC); }
-#ifdef CONFIG_COMPAT - COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp) { @@ -1574,7 +1572,6 @@ COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, current->restart_block.nanosleep.compat_rmtp = rmtp; return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC); } -#endif
/* * Functions related to boot-time initialization: diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c index b258bee13b02..b2dc97a39c35 100644 --- a/kernel/time/posix-stubs.c +++ b/kernel/time/posix-stubs.c @@ -148,6 +148,7 @@ COMPAT_SYS_NI(timer_settime); COMPAT_SYS_NI(timer_gettime); COMPAT_SYS_NI(getitimer); COMPAT_SYS_NI(setitimer); +#endif
COMPAT_SYSCALL_DEFINE2(clock_settime, const clockid_t, which_clock, struct compat_timespec __user *, tp) @@ -224,4 +225,3 @@ COMPAT_SYSCALL_DEFINE4(clock_nanosleep, clockid_t, which_clock, int, flags, HRTIMER_MODE_ABS : HRTIMER_MODE_REL, which_clock); } -#endif diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c index 13d6881f908b..0650ec27a673 100644 --- a/kernel/time/posix-timers.c +++ b/kernel/time/posix-timers.c @@ -1107,8 +1107,6 @@ SYSCALL_DEFINE2(clock_getres, const clockid_t, which_clock, return error; }
-#ifdef CONFIG_COMPAT - COMPAT_SYSCALL_DEFINE2(clock_settime, clockid_t, which_clock, struct compat_timespec __user *, tp) { @@ -1142,6 +1140,8 @@ COMPAT_SYSCALL_DEFINE2(clock_gettime, clockid_t, which_clock, return err; }
+#ifdef CONFIG_COMPAT + COMPAT_SYSCALL_DEFINE2(clock_adjtime, clockid_t, which_clock, struct compat_timex __user *, utp) { @@ -1166,6 +1166,8 @@ COMPAT_SYSCALL_DEFINE2(clock_adjtime, clockid_t, which_clock, return err; }
+#endif + COMPAT_SYSCALL_DEFINE2(clock_getres, clockid_t, which_clock, struct compat_timespec __user *, tp) { @@ -1183,8 +1185,6 @@ COMPAT_SYSCALL_DEFINE2(clock_getres, clockid_t, which_clock, return err; }
-#endif - /* * nanosleep for monotonic and realtime clocks */ @@ -1221,7 +1221,6 @@ SYSCALL_DEFINE4(clock_nanosleep, const clockid_t, which_clock, int, flags, return kc->nsleep(which_clock, flags, &t); }
-#ifdef CONFIG_COMPAT COMPAT_SYSCALL_DEFINE4(clock_nanosleep, clockid_t, which_clock, int, flags, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp) @@ -1246,7 +1245,6 @@ COMPAT_SYSCALL_DEFINE4(clock_nanosleep, clockid_t, which_clock, int, flags,
return kc->nsleep(which_clock, flags, &t); } -#endif
static const struct k_clock clock_realtime = { .clock_getres = posix_get_hrtimer_res,
On Fri, 10 Nov 2017, Deepa Dinamani wrote:
-#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp) { @@ -1574,7 +1572,6 @@ COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, current->restart_block.nanosleep.compat_rmtp = rmtp; return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC); } -#endif
So if I'm not missing something important this will make the compat syscall define available even for
CONFIG_X86_64=y CONFIG_IA32_EMULATION=n CONFIG_X86_X32=n
which is wrong because in that configuration we don't have any 32bit executable support. So why would we need a compat syscall in that case?
Thanks,
tglx
On Mon, Nov 13, 2017 at 1:03 PM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 10 Nov 2017, Deepa Dinamani wrote:
-#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp) { @@ -1574,7 +1572,6 @@ COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, current->restart_block.nanosleep.compat_rmtp = rmtp; return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC); } -#endif
So if I'm not missing something important this will make the compat syscall define available even for
CONFIG_X86_64=y CONFIG_IA32_EMULATION=n CONFIG_X86_X32=n
which is wrong because in that configuration we don't have any 32bit executable support. So why would we need a compat syscall in that case?
I was thinking that we would switch the meaning of CONFIG_64BIT_TIME after all architectures have enabled support for 64 bit time_t syscalls. After that, I was going to use the config to mean enable only 64 bit time_t support and so all compat syscalls would have #ifndef CONFIG_64BIT_TIME around them. Compat syscalls do not mean compat anymore for these time syscalls as the data structure is defined in a way that they are the same on 64 bit and 32 bit architectures.
But, you are right. It will leave in compat syscalls when it is not required now. I could add a dependency on CONFIG_COMPAT and __BITS_PER_LONG for now if you prefer. I could introduce additional dependencies later on.
-Deepa
On Tue, Nov 14, 2017 at 12:26 AM, Deepa Dinamani deepa.kernel@gmail.com wrote:
On Mon, Nov 13, 2017 at 1:03 PM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 10 Nov 2017, Deepa Dinamani wrote:
-#ifdef CONFIG_COMPAT
COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp) { @@ -1574,7 +1572,6 @@ COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, current->restart_block.nanosleep.compat_rmtp = rmtp; return hrtimer_nanosleep(&tu, HRTIMER_MODE_REL, CLOCK_MONOTONIC); } -#endif
So if I'm not missing something important this will make the compat syscall define available even for
CONFIG_X86_64=y CONFIG_IA32_EMULATION=n CONFIG_X86_X32=n
which is wrong because in that configuration we don't have any 32bit executable support. So why would we need a compat syscall in that case?
I was thinking that we would switch the meaning of CONFIG_64BIT_TIME after all architectures have enabled support for 64 bit time_t syscalls.
I think at that point the CONFIG_64BIT_TIME symbol can just go away, the name would not express any useful meaning any more once all architectures use 64-bit time types.
After that, I was going to use the config to mean enable only 64 bit time_t support and so all compat syscalls would have #ifndef CONFIG_64BIT_TIME around them. Compat syscalls do not mean compat anymore for these time syscalls as the data structure is defined in a way that they are the same on 64 bit and 32 bit architectures.
But, you are right. It will leave in compat syscalls when it is not required now. I could add a dependency on CONFIG_COMPAT and __BITS_PER_LONG for now if you prefer. I could introduce additional dependencies later on.
In kernel uapi headers, checking for __BITS_PER_LONG is appropriate, but we can't check for CONFIG_COMPAT there, since user space does not see the CONFIG_ namespace.
In the kernel, checking for 64-bit architectures is better done using the CONFIG_64BIT symbol.
So we could do
-#ifdef CONFIG_COMPAT +#if defined(CONFIG_COMPAT) || !defined(CONFIG_64BIT) - COMPAT_SYSCALL_DEFINE2(nanosleep, struct compat_timespec __user *, rqtp, struct compat_timespec __user *, rmtp)
The trick with renaming the Kconfig symbol worked better with the old CONFIG_COMPAT_TIME name that I had in an earlier draft of the series, it would also enable us to leave out the 32-bit time compat syscall implementation for future architectures that never had the old uapi.
Arnd
From: Arnd Bergmann arnd@arndb.de
There are a total of 53 system calls (aside from ioctl) that pass a time_t or derived data structure as an argument, and in order to extend time_t to 64-bit, we have to replace them with new system calls and keep providing backwards compatibility.
To avoid adding completely new and untested code for this purpose, we introduce a new CONFIG_64BIT_TIME symbol. Every architecture that supports new 64 bit time_t syscalls enables this config via ARCH_HAS_64BIT_TIME.
After this is done for all architectures, the CONFIG_64BIT_TIME symbol can be made a user-selected option, to enable users to build a kernel that only provides y2038-safe system calls by making 32 time_t syscalls conditionally included based on the above config.
Signed-off-by: Arnd Bergmann arnd@arndb.de Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- arch/Kconfig | 11 +++++++++++ 1 file changed, 11 insertions(+)
diff --git a/arch/Kconfig b/arch/Kconfig index 8911ff37335a..3266ac1a4ff7 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -875,6 +875,17 @@ config OLD_SIGACTION config COMPAT_OLD_SIGACTION bool
+config ARCH_HAS_64BIT_TIME + def_bool n + +config CONFIG_64BIT_TIME + def_bool ARCH_HAS_64BIT_TIME + help + This should be selected by all architectures that need to support + new system calls with a 64-bit time_t. This is relevant on all 32-bit + architectures, and 64-bit architectures as part of compat syscall + handling. + config ARCH_NO_COHERENT_DMA_MMAP bool
On Fri, 2017-11-10 at 14:42 -0800, Deepa Dinamani wrote:
From: Arnd Bergmann arnd@arndb.de
There are a total of 53 system calls (aside from ioctl) that pass a time_t or derived data structure as an argument, and in order to extend time_t to 64-bit, we have to replace them with new system calls and keep providing backwards compatibility.
To avoid adding completely new and untested code for this purpose, we introduce a new CONFIG_64BIT_TIME symbol. Every architecture that supports new 64 bit time_t syscalls enables this config via ARCH_HAS_64BIT_TIME.
After this is done for all architectures, the CONFIG_64BIT_TIME symbol can be made a user-selected option, to enable users to build a kernel that only provides y2038-safe system calls by making 32 time_t syscalls conditionally included based on the above config.
I don't understand why we would want to change the semantics of CONFIG_64BIT_TIME symbol from "enable 64-bit time support" to "disable 32-bit time support".
Why not add two config symbols:
config 32BIT_TIME def_bool COMPAT || !64BIT
config 64BIT_TIME def_bool ARCH_HAS_64BIT_TIME
and then make 32BIT_TIME user-configurable later?
Signed-off-by: Arnd Bergmann arnd@arndb.de Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com
arch/Kconfig | 11 +++++++++++ Â 1 file changed, 11 insertions(+)
diff --git a/arch/Kconfig b/arch/Kconfig index 8911ff37335a..3266ac1a4ff7 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -875,6 +875,17 @@ config OLD_SIGACTION  config COMPAT_OLD_SIGACTION  bool  +config ARCH_HAS_64BIT_TIME
- def_bool n
+config CONFIG_64BIT_TIME
The CONFIG_ prefix is added by kconfig scripts and shouldn't be used in the Kconfig file.
Ben.
- def_bool ARCH_HAS_64BIT_TIME
- help
- This should be selected by all architectures that need to support
- new system calls with a 64-bit time_t. This is relevant on all 32-bit
- architectures, and 64-bit architectures as part of compat syscall
- handling.
config ARCH_NO_COHERENT_DMA_MMAP Â bool
On Fri, Dec 8, 2017 at 8:53 AM, Ben Hutchings ben.hutchings@codethink.co.uk wrote:
On Fri, 2017-11-10 at 14:42 -0800, Deepa Dinamani wrote:
From: Arnd Bergmann arnd@arndb.de
There are a total of 53 system calls (aside from ioctl) that pass a time_t or derived data structure as an argument, and in order to extend time_t to 64-bit, we have to replace them with new system calls and keep providing backwards compatibility.
To avoid adding completely new and untested code for this purpose, we introduce a new CONFIG_64BIT_TIME symbol. Every architecture that supports new 64 bit time_t syscalls enables this config via ARCH_HAS_64BIT_TIME.
After this is done for all architectures, the CONFIG_64BIT_TIME symbol can be made a user-selected option, to enable users to build a kernel that only provides y2038-safe system calls by making 32 time_t syscalls conditionally included based on the above config.
I don't understand why we would want to change the semantics of CONFIG_64BIT_TIME symbol from "enable 64-bit time support" to "disable 32-bit time support".
Why not add two config symbols:
config 32BIT_TIME def_bool COMPAT || !64BIT
config 64BIT_TIME def_bool ARCH_HAS_64BIT_TIME
and then make 32BIT_TIME user-configurable later?
This was already discussed on the review and we have an updated version:
https://lkml.org/lkml/2017/11/27/938
Signed-off-by: Arnd Bergmann arnd@arndb.de Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com
arch/Kconfig | 11 +++++++++++ 1 file changed, 11 insertions(+)
diff --git a/arch/Kconfig b/arch/Kconfig index 8911ff37335a..3266ac1a4ff7 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -875,6 +875,17 @@ config OLD_SIGACTION config COMPAT_OLD_SIGACTION bool
+config ARCH_HAS_64BIT_TIME
def_bool n+config CONFIG_64BIT_TIME
The CONFIG_ prefix is added by kconfig scripts and shouldn't be used in the Kconfig file.
Yes, this was a typo and was fixed by the next revision which has already been posted:
https://lkml.org/lkml/2017/11/27/938
-Deepa
On Fri, 2017-12-08 at 09:01 -0800, Deepa Dinamani wrote:
On Fri, Dec 8, 2017 at 8:53 AM, Ben Hutchings
ben.hutchings@codethink.co.uk wrote: On Fri, 2017-11-10 at 14:42 -0800, Deepa Dinamani wrote:
From: Arnd Bergmann arnd@arndb.de
There are a total of 53 system calls (aside from ioctl) that pass a time_t or derived data structure as an argument, and in order to extend time_t to 64-bit, we have to replace them with new system calls and keep providing backwards compatibility.
To avoid adding completely new and untested code for this purpose, we introduce a new CONFIG_64BIT_TIME symbol. Every architecture that supports new 64 bit time_t syscalls enables this config via ARCH_HAS_64BIT_TIME.
After this is done for all architectures, the CONFIG_64BIT_TIME symbol can be made a user-selected option, to enable users to build a kernel that only provides y2038-safe system calls by making 32 time_t syscalls conditionally included based on the above config.
I don't understand why we would want to change the semantics of CONFIG_64BIT_TIME symbol from "enable 64-bit time support" to "disable 32-bit time support".
Why not add two config symbols:
config 32BIT_TIME Â Â Â Â Â Â Â Â def_bool COMPAT || !64BIT
config 64BIT_TIME Â Â Â Â Â Â Â Â def_bool ARCH_HAS_64BIT_TIME
and then make 32BIT_TIME user-configurable later?
This was already discussed on the review and we have an updated version:
Sorry, I'll move on to reviewing that.
Ben.
The new struct __kernel_timespec is similar to current internal kernel struct timespec64 on 64 bit architecture. The compat structure however is similar to below on little endian systems (padding and tv_nsec are switched for big endian systems):
typedef s32 compat_long_t; typedef s64 compat_kernel_time64_t;
struct compat_kernel_timespec { compat_kernel_time64_t tv_sec; compat_long_t tv_nsec; compat_long_t padding; };
This allows for both the native and compat representations to be the same and syscalls using this type as part of their ABI can have a single entry point to both.
Note that the compat define is not included anywhere in the kernel explicitly to avoid confusion.
These types will be used by the new syscalls that will be introduced in the consequent patches. Most of the new syscalls are just an update to the existing native ones with this new type. Hence, put this new type under an ifdef so that the architectures can define CONFIG_64BIT_TIME when they are ready to handle this switch.
Cc: linux-arch@vger.kernel.org Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- include/linux/time64.h | 10 +++++++++- include/uapi/asm-generic/posix_types.h | 1 + include/uapi/linux/time.h | 7 +++++++ 3 files changed, 17 insertions(+), 1 deletion(-)
diff --git a/include/linux/time64.h b/include/linux/time64.h index 93d39499838e..0d96887ba4e0 100644 --- a/include/linux/time64.h +++ b/include/linux/time64.h @@ -2,12 +2,20 @@ #ifndef _LINUX_TIME64_H #define _LINUX_TIME64_H
-#include <uapi/linux/time.h> #include <linux/math64.h>
typedef __s64 time64_t; typedef __u64 timeu64_t;
+/* CONFIG_64BIT_TIME enables new 64 bit time_t syscalls in the compat path + * and 32-bit emulation. + */ +#ifndef CONFIG_64BIT_TIME +#define __kernel_timespec timespec +#endif + +#include <uapi/linux/time.h> + #if __BITS_PER_LONG == 64 /* this trick allows us to optimize out timespec64_to_timespec */ # define timespec64 timespec diff --git a/include/uapi/asm-generic/posix_types.h b/include/uapi/asm-generic/posix_types.h index 5e6ea22bd525..f0733a26ebfc 100644 --- a/include/uapi/asm-generic/posix_types.h +++ b/include/uapi/asm-generic/posix_types.h @@ -87,6 +87,7 @@ typedef struct { typedef __kernel_long_t __kernel_off_t; typedef long long __kernel_loff_t; typedef __kernel_long_t __kernel_time_t; +typedef long long __kernel_time64_t; typedef __kernel_long_t __kernel_clock_t; typedef int __kernel_timer_t; typedef int __kernel_clockid_t; diff --git a/include/uapi/linux/time.h b/include/uapi/linux/time.h index 53f8dd84beb5..0ad4510884b0 100644 --- a/include/uapi/linux/time.h +++ b/include/uapi/linux/time.h @@ -42,6 +42,13 @@ struct itimerval { struct timeval it_value; /* current value */ };
+#ifndef __kernel_timespec +struct __kernel_timespec { + __kernel_time64_t tv_sec; /* seconds */ + long long tv_nsec; /* nanoseconds */ +}; +#endif + /* * The IDs of the various system clocks (for POSIX.1b interval timers): */
get/put_timespec64() interfaces will eventually be used for conversions between the new y2038 safe struct __kernel_timespec and struct timespec64.
The new y2038 safe syscalls have a common entry for native and compat interfaces. On compat interfaces, the high order bits of nanoseconds are should to be zeroed out. This is because the application code or the libc do not garuntee zeroing of these. If used without zeroing, kernel might be at risk of using timespec values incorrectly.
Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- include/linux/time.h | 4 ++-- kernel/time/time.c | 10 ++++++++-- 2 files changed, 10 insertions(+), 4 deletions(-)
diff --git a/include/linux/time.h b/include/linux/time.h index 4b62a2c0a661..aed74463592d 100644 --- a/include/linux/time.h +++ b/include/linux/time.h @@ -10,9 +10,9 @@ extern struct timezone sys_tz;
int get_timespec64(struct timespec64 *ts, - const struct timespec __user *uts); + const struct __kernel_timespec __user *uts); int put_timespec64(const struct timespec64 *ts, - struct timespec __user *uts); + struct __kernel_timespec __user *uts); int get_itimerspec64(struct itimerspec64 *it, const struct itimerspec __user *uit); int put_itimerspec64(const struct itimerspec64 *it, diff --git a/kernel/time/time.c b/kernel/time/time.c index bd4e6c7dd689..476fd00c9bb0 100644 --- a/kernel/time/time.c +++ b/kernel/time/time.c @@ -841,7 +841,7 @@ struct timespec64 timespec64_add_safe(const struct timespec64 lhs, }
int get_timespec64(struct timespec64 *ts, - const struct timespec __user *uts) + const struct __kernel_timespec __user *uts) { struct timespec kts; int ret; @@ -851,6 +851,11 @@ int get_timespec64(struct timespec64 *ts, return -EFAULT;
ts->tv_sec = kts.tv_sec; + + /* Zero out the padding for 32 bit systems or in compat mode */ + if (!IS_ENABLED(CONFIG_64BIT) || in_compat_syscall()) + kts.tv_nsec &= 0xFFFFFFFFUL; + ts->tv_nsec = kts.tv_nsec;
return 0; @@ -858,12 +863,13 @@ int get_timespec64(struct timespec64 *ts, EXPORT_SYMBOL_GPL(get_timespec64);
int put_timespec64(const struct timespec64 *ts, - struct timespec __user *uts) + struct __kernel_timespec __user *uts) { struct timespec kts = { .tv_sec = ts->tv_sec, .tv_nsec = ts->tv_nsec }; + return copy_to_user(uts, &kts, sizeof(kts)) ? -EFAULT : 0; } EXPORT_SYMBOL_GPL(put_timespec64);
On Fri, Nov 10, 2017 at 11:42 PM, Deepa Dinamani deepa.kernel@gmail.com wrote:
/* Zero out the padding for 32 bit systems or in compat mode */if (!IS_ENABLED(CONFIG_64BIT) || in_compat_syscall())kts.tv_nsec &= 0xFFFFFFFFUL;
Is this correct for x32? Since x32 has a 64-bit tv_nsec, I think we currently return an error in timespec64_valid when e.g. you pass 0xffffffff00000000 as the tv_nsec value, while after this patch x32 would behave the same way as other 32-bit architectures and silently clear the upper bits.
Arnd
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 b2dc97a39c35..684d470dba2c 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 0650ec27a673..d482359bce1b 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;
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@vger.kernel.org Signed-off-by: Deepa Dinamani deepa.kernel@gmail.com --- include/linux/restart_block.h | 7 ++----- include/linux/syscalls.h | 6 +++--- kernel/time/hrtimer.c | 4 ++-- kernel/time/posix-stubs.c | 4 ++-- kernel/time/posix-timers.c | 4 ++-- 5 files changed, 11 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 e7db66af5a23..ce6cbb687c47 100644 --- a/kernel/time/hrtimer.c +++ b/kernel/time/hrtimer.c @@ -1541,8 +1541,8 @@ long hrtimer_nanosleep(const struct timespec64 *rqtp, return ret; }
-SYSCALL_DEFINE2(nanosleep, struct timespec __user *, rqtp, - struct timespec __user *, rmtp) +SYSCALL_DEFINE2(nanosleep, struct __kernel_timespec __user *, rqtp, + struct __kernel_timespec __user *, rmtp) { struct timespec64 tu;
diff --git a/kernel/time/posix-stubs.c b/kernel/time/posix-stubs.c index 684d470dba2c..cdc5fed5f89b 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 d482359bce1b..b8e1bf1aa00f 100644 --- a/kernel/time/posix-timers.c +++ b/kernel/time/posix-timers.c @@ -1197,8 +1197,8 @@ 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;
On Fri, Nov 10, 2017 at 11:42 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
Big picture is as per the lwn article: https://lwn.net/Articles/643234/
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:
- Enable compat syscalls unconditionally.
- Add a new __kernel_timespec type to be used as the data structure for all the new syscalls.
- 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.
Arnd Bergmann (1): y2038: introduce CONFIG_64BIT_TIME
The series looks good to me overall, and I've added it to my build-testing tree to see if it shows any regressions in random configurations.
I had on concern about x32, maybe we should check for "COMPAT_USE_64BIT_TIME" before zeroing out the tv_nsec bits.
Regarding CONFIG_COMPAT_TIME/CONFIG_64BIT_TIME, would it help to just leave out that part for now and unconditionally define '__kernel_timespec' as 'timespec' until we are ready to convert the architectures?
Arnd
I had on concern about x32, maybe we should check for "COMPAT_USE_64BIT_TIME" before zeroing out the tv_nsec bits.
Thanks, I think you are right. I had the check conditional on CONFIG_64BIT_TIME and then removed as I forgot why I added it. :)
Regarding CONFIG_COMPAT_TIME/CONFIG_64BIT_TIME, would it help to just leave out that part for now and unconditionally define '__kernel_timespec' as 'timespec' until we are ready to convert the architectures?
Another approach would be to use separate configs:
1. To indicate 64 bit time_t syscall support. This will be dependent on architectures as CONFIG_64BIT_TIME. We can delete this once all architectures have provided support for this.
2. Another config (maybe COMPAT_32BIT_TIME?) to be introduced later, which will compile out all syscalls/ features that use 32 bit time_t. This can help build a y2038 safe kernel later.
Would this work for everyone?
-Deepa
On Wed, 15 Nov 2017, Deepa Dinamani wrote:
I had on concern about x32, maybe we should check for "COMPAT_USE_64BIT_TIME" before zeroing out the tv_nsec bits.
Thanks, I think you are right. I had the check conditional on CONFIG_64BIT_TIME and then removed as I forgot why I added it. :)
Regarding CONFIG_COMPAT_TIME/CONFIG_64BIT_TIME, would it help to just leave out that part for now and unconditionally define '__kernel_timespec' as 'timespec' until we are ready to convert the architectures?
Another approach would be to use separate configs:
- To indicate 64 bit time_t syscall support. This will be dependent
on architectures as CONFIG_64BIT_TIME. We can delete this once all architectures have provided support for this.
- Another config (maybe COMPAT_32BIT_TIME?) to be introduced later,
which will compile out all syscalls/ features that use 32 bit time_t. This can help build a y2038 safe kernel later.
Would this work for everyone?
Having extra config switches which are selectable by architectures and removed when everything is converted is definitely the right way to go.
That allows you to gradually convert stuff w/o inflicting wreckage all over the place.
Thanks,
tglx
On Thu, Nov 16, 2017 at 10:04 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Wed, 15 Nov 2017, Deepa Dinamani wrote:
I had on concern about x32, maybe we should check for "COMPAT_USE_64BIT_TIME" before zeroing out the tv_nsec bits.
Thanks, I think you are right. I had the check conditional on CONFIG_64BIT_TIME and then removed as I forgot why I added it. :)
Regarding CONFIG_COMPAT_TIME/CONFIG_64BIT_TIME, would it help to just leave out that part for now and unconditionally define '__kernel_timespec' as 'timespec' until we are ready to convert the architectures?
Another approach would be to use separate configs:
- To indicate 64 bit time_t syscall support. This will be dependent
on architectures as CONFIG_64BIT_TIME. We can delete this once all architectures have provided support for this.
- Another config (maybe COMPAT_32BIT_TIME?) to be introduced later,
which will compile out all syscalls/ features that use 32 bit time_t. This can help build a y2038 safe kernel later.
Would this work for everyone?
Having extra config switches which are selectable by architectures and removed when everything is converted is definitely the right way to go.
That allows you to gradually convert stuff w/o inflicting wreckage all over the place.
The CONFIG_64BIT_TIME would do that nicely for the new stuff like the conditional definition of __kernel_timespec, this one would get removed after we convert all architectures.
A second issue is how to control the compilation of the compat syscalls. CONFIG_COMPAT_32BIT_TIME handles that and could be defined in Kconfig as 'def_bool (!64BIT && CONFIG_64BIT_TIME) || COMPAT', this is then just a more readable way of expressing exactly when the functions should be built.
For completeness, there may be a third category, depending on how we handle things like sys_nanosleep(): Here, we want the native sys_nanosleep on 64-bit architectures, and compat_sys_nanosleep() to handle the 32-bit time_t variant on both 32-bit and 64-bit targets, but our plan is to not have a native 32-bit sys_nanosleep on 32-bit architectures any more, as new glibc should call clock_nanosleep() with a new syscall number instead. Should we then enclose sys_nanosleep in "#if !defined(CONFIG_64BIT_TIME) || defined(CONFIG_64BIT)", or should we try to come up with another Kconfig symbol name that expresses this better?
Arnd
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Thu, Nov 16, 2017 at 10:04 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Wed, 15 Nov 2017, Deepa Dinamani wrote:
Would this work for everyone?
Having extra config switches which are selectable by architectures and removed when everything is converted is definitely the right way to go.
That allows you to gradually convert stuff w/o inflicting wreckage all over the place.
The CONFIG_64BIT_TIME would do that nicely for the new stuff like the conditional definition of __kernel_timespec, this one would get removed after we convert all architectures.
A second issue is how to control the compilation of the compat syscalls. CONFIG_COMPAT_32BIT_TIME handles that and could be defined in Kconfig as 'def_bool (!64BIT && CONFIG_64BIT_TIME) || COMPAT', this is then just a more readable way of expressing exactly when the functions should be built.
For completeness, there may be a third category, depending on how we handle things like sys_nanosleep(): Here, we want the native sys_nanosleep on 64-bit architectures, and compat_sys_nanosleep() to handle the 32-bit time_t variant on both 32-bit and 64-bit targets, but our plan is to not have a native 32-bit sys_nanosleep on 32-bit architectures any more, as new glibc should call clock_nanosleep() with a new syscall number instead. Should we then enclose
Isn't that going to break existing userspace?
Thanks
tglx
On Fri, Nov 17, 2017 at 9:58 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Thu, Nov 16, 2017 at 10:04 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Wed, 15 Nov 2017, Deepa Dinamani wrote:
Would this work for everyone?
Having extra config switches which are selectable by architectures and removed when everything is converted is definitely the right way to go.
That allows you to gradually convert stuff w/o inflicting wreckage all over the place.
The CONFIG_64BIT_TIME would do that nicely for the new stuff like the conditional definition of __kernel_timespec, this one would get removed after we convert all architectures.
A second issue is how to control the compilation of the compat syscalls. CONFIG_COMPAT_32BIT_TIME handles that and could be defined in Kconfig as 'def_bool (!64BIT && CONFIG_64BIT_TIME) || COMPAT', this is then just a more readable way of expressing exactly when the functions should be built.
For completeness, there may be a third category, depending on how we handle things like sys_nanosleep(): Here, we want the native sys_nanosleep on 64-bit architectures, and compat_sys_nanosleep() to handle the 32-bit time_t variant on both 32-bit and 64-bit targets, but our plan is to not have a native 32-bit sys_nanosleep on 32-bit architectures any more, as new glibc should call clock_nanosleep() with a new syscall number instead. Should we then enclose
Isn't that going to break existing userspace?
No, syscall that existing 32-bit user space enters would be handled by compat_sys_nanosleep() on both 32-bit and 64-bit kernels at that point. The idea here is to make the code path more uniform between 32-bit and 64-bit kernels.
Arnd
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 9:58 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Thu, Nov 16, 2017 at 10:04 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Wed, 15 Nov 2017, Deepa Dinamani wrote:
Would this work for everyone?
Having extra config switches which are selectable by architectures and removed when everything is converted is definitely the right way to go.
That allows you to gradually convert stuff w/o inflicting wreckage all over the place.
The CONFIG_64BIT_TIME would do that nicely for the new stuff like the conditional definition of __kernel_timespec, this one would get removed after we convert all architectures.
A second issue is how to control the compilation of the compat syscalls. CONFIG_COMPAT_32BIT_TIME handles that and could be defined in Kconfig as 'def_bool (!64BIT && CONFIG_64BIT_TIME) || COMPAT', this is then just a more readable way of expressing exactly when the functions should be built.
For completeness, there may be a third category, depending on how we handle things like sys_nanosleep(): Here, we want the native sys_nanosleep on 64-bit architectures, and compat_sys_nanosleep() to handle the 32-bit time_t variant on both 32-bit and 64-bit targets, but our plan is to not have a native 32-bit sys_nanosleep on 32-bit architectures any more, as new glibc should call clock_nanosleep() with a new syscall number instead. Should we then enclose
Isn't that going to break existing userspace?
No, syscall that existing 32-bit user space enters would be handled by compat_sys_nanosleep() on both 32-bit and 64-bit kernels at that point. The idea here is to make the code path more uniform between 32-bit and 64-bit kernels.
So on a 32bit system compat_sys_nanosleep() would be the legacy sys_nanosleep() with the existing syscall number, but you don't want to introduce a new sys_nanosleep64() for 32bit. That makes a lot of sense.
So back to your original question whether to use #if (MAGIC logic) or a separate config symbol. Please use the latter, these magic logic constructs are harder to read and prone to get wrong at some point. Having the decision logic in one place is always the right thing to do.
Thanks,
tglx
On Fri, Nov 17, 2017 at 10:54 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 9:58 AM, Thomas Gleixner tglx@linutronix.de wrote:
No, syscall that existing 32-bit user space enters would be handled by compat_sys_nanosleep() on both 32-bit and 64-bit kernels at that point. The idea here is to make the code path more uniform between 32-bit and 64-bit kernels.
So on a 32bit system compat_sys_nanosleep() would be the legacy sys_nanosleep() with the existing syscall number, but you don't want to introduce a new sys_nanosleep64() for 32bit. That makes a lot of sense.
So back to your original question whether to use #if (MAGIC logic) or a separate config symbol. Please use the latter, these magic logic constructs are harder to read and prone to get wrong at some point. Having the decision logic in one place is always the right thing to do.
How about this:
config LEGACY_TIME_SYSCALLS def_bool 64BIT || !64BIT_TIME help 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 disable this.
Arnd
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 10:54 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 9:58 AM, Thomas Gleixner tglx@linutronix.de wrote:
No, syscall that existing 32-bit user space enters would be handled by compat_sys_nanosleep() on both 32-bit and 64-bit kernels at that point. The idea here is to make the code path more uniform between 32-bit and 64-bit kernels.
So on a 32bit system compat_sys_nanosleep() would be the legacy sys_nanosleep() with the existing syscall number, but you don't want to introduce a new sys_nanosleep64() for 32bit. That makes a lot of sense.
So back to your original question whether to use #if (MAGIC logic) or a separate config symbol. Please use the latter, these magic logic constructs are harder to read and prone to get wrong at some point. Having the decision logic in one place is always the right thing to do.
How about this:
config LEGACY_TIME_SYSCALLS def_bool 64BIT || !64BIT_TIME help 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.
s/handles/handling/ ????
New architectures should not explicitly disable this.
New architectures should never enable this, right?
Thanks,
tglx
On Fri, Nov 17, 2017 at 11:40 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 10:54 AM, Thomas Gleixner tglx@linutronix.de wrote:
On Fri, 17 Nov 2017, Arnd Bergmann wrote:
On Fri, Nov 17, 2017 at 9:58 AM, Thomas Gleixner tglx@linutronix.de wrote:
No, syscall that existing 32-bit user space enters would be handled by compat_sys_nanosleep() on both 32-bit and 64-bit kernels at that point. The idea here is to make the code path more uniform between 32-bit and 64-bit kernels.
So on a 32bit system compat_sys_nanosleep() would be the legacy sys_nanosleep() with the existing syscall number, but you don't want to introduce a new sys_nanosleep64() for 32bit. That makes a lot of sense.
So back to your original question whether to use #if (MAGIC logic) or a separate config symbol. Please use the latter, these magic logic constructs are harder to read and prone to get wrong at some point. Having the decision logic in one place is always the right thing to do.
How about this:
config LEGACY_TIME_SYSCALLS def_bool 64BIT || !64BIT_TIME help 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.
s/handles/handling/ ????
I meant "handlers".
New architectures should not explicitly disable this.New architectures should never enable this, right?
Right, I got an extra "not". I guess if Deepa incorporates the new option, she can also improve my English ;-)
Arnd
-
Arnd Bergmann -
Ben Hutchings -
Deepa Dinamani -
Steven Rostedt -
Thomas Gleixner