Re: [PATCH] linux-user: Fix 'clock_nanosleep()' implementation

[Laurent Vivier <laurent@vivier.eu>]

Le 27/07/2020 à 22:13, Filip Bozuta a écrit :
> Implementation of syscall 'clock_nanosleep()' in 'syscall.c' uses
> functions 'target_to_host_timespec()' and 'host_to_target_timespec()'
> to transfer the value of 'struct timespec' between target and host.
> However, the implementation doesn't check whether this conversion
> succeeds and thus can return an unaproppriate error instead of 'EFAULT'
> that is expected. This was confirmed with the modified LTP test suite
> where testcases with bad 'struct timespec' adress for 'clock_nanosleep()'
> were added. This modified LTP suite can be found at:
> https://github.com/bozutaf/ltp
> 
> (Patch with this new test case will be sent to LTP mailing list soon)
> 
> Signed-off-by: Filip Bozuta <Filip.Bozuta@syrmia.com>
> ---
>  linux-user/syscall.c | 9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/linux-user/syscall.c b/linux-user/syscall.c
> index f5c4f6b95d..9f06dde947 100644
> --- a/linux-user/syscall.c
> +++ b/linux-user/syscall.c
> @@ -11828,7 +11828,9 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1,
>      case TARGET_NR_clock_nanosleep:
>      {
>          struct timespec ts;
> -        target_to_host_timespec(&ts, arg3);
> +        if (target_to_host_timespec(&ts, arg3)) {
> +            return -TARGET_EFAULT;
> +        }
>          ret = get_errno(safe_clock_nanosleep(arg1, arg2,
>                                               &ts, arg4 ? &ts : NULL));
>          /*
> @@ -11836,8 +11838,9 @@ static abi_long do_syscall1(void *cpu_env, int num, abi_long arg1,
>           * with error -TARGET_EINTR and if arg4 is not NULL and arg2 is not
>           * TIMER_ABSTIME, it returns the remaining unslept time in arg4.
>           */
> -        if (ret == -TARGET_EINTR && arg4 && arg2 != TIMER_ABSTIME) {
> -            host_to_target_timespec(arg4, &ts);
> +        if (ret == -TARGET_EINTR && arg4 && arg2 != TIMER_ABSTIME &&
> +            host_to_target_timespec(arg4, &ts)) {
> +              return -TARGET_EFAULT;
>          }
>  
>          return ret;
> 

Applied to my linux-user-for-5.2 branch.

Thanks,
Laurent