On Tue, Jan 09, 2018 at 01:56:16PM +0100, Willy Tarreau wrote:
> This allows to report the current state of the PTI protection and to
> enable or disable it for the current process. The state change is only
> allowed if the mm is not shared (no threads have been created yet).
>
> Setting the flag to disable the protection is subject to CAP_SYS_RAWIO.
> However it is possible to re-enable the protection without this privilege.
>
> Signed-off-by: Willy Tarreau <[email protected]>
> Cc: Andy Lutomirski <[email protected]>
> Cc: Borislav Petkov <[email protected]>
> Cc: Brian Gerst <[email protected]>
> Cc: Dave Hansen <[email protected]>
> Cc: Ingo Molnar <[email protected]>
> Cc: Linus Torvalds <[email protected]>
> Cc: Peter Zijlstra <[email protected]>
> Cc: Thomas Gleixner <[email protected]>
> Cc: Josh Poimboeuf <[email protected]>
> Cc: "H. Peter Anvin" <[email protected]>
> Cc: Kees Cook <[email protected]>
>
> v2:
> - use {set,clear}_thread_flag() as recommended by Peter
> - use task->mm->context.pti_disable instead of task flag
> - check for mm_users == 1
> - check for CAP_SYS_RAWIO only when setting, not clearing
> - make the code depend on CONFIG_PAGE_TABLE_ISOLATION
> ---
> arch/x86/include/uapi/asm/prctl.h | 3 +++
> arch/x86/kernel/process_64.c | 15 +++++++++++++++
> 2 files changed, 18 insertions(+)
>
> diff --git a/arch/x86/include/uapi/asm/prctl.h
> b/arch/x86/include/uapi/asm/prctl.h
> index 5a6aac9..1f1b5bc 100644
> --- a/arch/x86/include/uapi/asm/prctl.h
> +++ b/arch/x86/include/uapi/asm/prctl.h
> @@ -10,6 +10,9 @@
> #define ARCH_GET_CPUID 0x1011
> #define ARCH_SET_CPUID 0x1012
>
> +#define ARCH_GET_NOPTI 0x1021
> +#define ARCH_SET_NOPTI 0x1022
> +
> #define ARCH_MAP_VDSO_X32 0x2001
> #define ARCH_MAP_VDSO_32 0x2002
> #define ARCH_MAP_VDSO_64 0x2003
> diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
> index c754662..9516310 100644
> --- a/arch/x86/kernel/process_64.c
> +++ b/arch/x86/kernel/process_64.c
> @@ -654,7 +654,22 @@ long do_arch_prctl_64(struct task_struct *task, int
> option, unsigned long arg2)
> ret = put_user(base, (unsigned long __user *)arg2);
> break;
> }
> +#ifdef CONFIG_PAGE_TABLE_ISOLATION
Actually, I meant to add a new CONFIG item only for this feature which
depends on CONFIG_PAGE_TABLE_ISOLATION. So that people can disable the
per-mm thing when they don't want it.
Thx.
--
Regards/Gruss,
Boris.
Good mailing practices for 400: avoid top-posting and trim the reply.
