On compat interfaces, the high order bits of nanoseconds should be zeroed out. This is because the application code or the libc do not guarantee zeroing of these. If used without zeroing, kernel might be at risk of using timespec values incorrectly.
Originally it was handled correctly, but lost during is_compat_syscall() cleanup. Revert the condition back to check CONFIG_64BIT.
Reported-by: Ben Hutchings ben.hutchings@codethink.co.uk Fixes: 98f76206b335 ("compat: Cleanup in_compat_syscall() callers") Cc: Arnd Bergmann arnd@arndb.de Cc: Ben Hutchings ben.hutchings@codethink.co.uk Cc: John Stultz john.stultz@linaro.org Cc: Stephen Boyd sboyd@kernel.org Cc: Thomas Gleixner tglx@linutronix.de Cc: y2038@lists.linaro.org Cc: stable@vger.kernel.org # v4.20+ Signed-off-by: Dmitry Safonov dima@arista.com --- kernel/time/time.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/kernel/time/time.c b/kernel/time/time.c index 5c54ca632d08..1cb045c5c97e 100644 --- a/kernel/time/time.c +++ b/kernel/time/time.c @@ -881,7 +881,7 @@ int get_timespec64(struct timespec64 *ts, ts->tv_sec = kts.tv_sec;
/* Zero out the padding for 32 bit systems or in compat mode */ - if (IS_ENABLED(CONFIG_64BIT_TIME) && in_compat_syscall()) + if (IS_ENABLED(CONFIG_64BIT_TIME) && (!IS_ENABLED(CONFIG_64BIT) || in_compat_syscall())) kts.tv_nsec &= 0xFFFFFFFFUL;
ts->tv_nsec = kts.tv_nsec;
The following commit has been merged into the timers/urgent branch of tip:
Commit-ID: 7b8474466ed97be458c825f34a85f2c2b84c3f95 Gitweb: https://git.kernel.org/tip/7b8474466ed97be458c825f34a85f2c2b84c3f95 Author: Dmitry Safonov dima@arista.com AuthorDate: Thu, 21 Nov 2019 00:03:03 Committer: Thomas Gleixner tglx@linutronix.de CommitterDate: Thu, 21 Nov 2019 01:17:58 +01:00
time: Zero the upper 32-bits in __kernel_timespec on 32-bit
On compat interfaces, the high order bits of nanoseconds should be zeroed out. This is because the application code or the libc do not guarantee zeroing of these. If used without zeroing, kernel might be at risk of using timespec values incorrectly.
Originally it was handled correctly, but lost during is_compat_syscall() cleanup. Revert the condition back to check CONFIG_64BIT.
Fixes: 98f76206b335 ("compat: Cleanup in_compat_syscall() callers") Reported-by: Ben Hutchings ben.hutchings@codethink.co.uk Signed-off-by: Dmitry Safonov dima@arista.com Signed-off-by: Thomas Gleixner tglx@linutronix.de Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20191121000303.126523-1-dima@arista.com --- kernel/time/time.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/kernel/time/time.c b/kernel/time/time.c index 5c54ca6..83f403e 100644 --- a/kernel/time/time.c +++ b/kernel/time/time.c @@ -881,7 +881,8 @@ int get_timespec64(struct timespec64 *ts, ts->tv_sec = kts.tv_sec;
/* Zero out the padding for 32 bit systems or in compat mode */ - if (IS_ENABLED(CONFIG_64BIT_TIME) && in_compat_syscall()) + if (IS_ENABLED(CONFIG_64BIT_TIME) && (!IS_ENABLED(CONFIG_64BIT) || + in_compat_syscall())) kts.tv_nsec &= 0xFFFFFFFFUL;
ts->tv_nsec = kts.tv_nsec;
On Thu, Nov 21, 2019 at 1:03 AM Dmitry Safonov dima@arista.com wrote:
On compat interfaces, the high order bits of nanoseconds should be zeroed out. This is because the application code or the libc do not guarantee zeroing of these. If used without zeroing, kernel might be at risk of using timespec values incorrectly.
Originally it was handled correctly, but lost during is_compat_syscall() cleanup. Revert the condition back to check CONFIG_64BIT.
Reported-by: Ben Hutchings ben.hutchings@codethink.co.uk Fixes: 98f76206b335 ("compat: Cleanup in_compat_syscall() callers")
Reviewed-by: Arnd Bergmann arnd@arndb.de
linux-stable-mirror@lists.linaro.org
-
Arnd Bergmann -
Dmitry Safonov -
tip-bot2 for Dmitry Safonov