Re: [PATCH 01/23] x86/fpu: Use ULL for shift in xfeature_uncompacted_offset()

From: Borislav Petkov
Date: Thu Nov 08 2018 - 06:38:28 EST


On Wed, Nov 07, 2018 at 08:48:36PM +0100, Sebastian Andrzej Siewior wrote:
> The xfeature mask is 64bit so a shift from a number to its mask should
> have LL prefix or else nr > 31 will be lost. This is not a problem now
> but should XFEATURE_MASK_SUPERVISOR gain a bit >31 then this check won't
> catch it.
>
> Add a ULL suffix so >31bit number will be properly shifted.
>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
> ---
> arch/x86/kernel/fpu/xstate.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
> index 87a57b7642d36..9dc0a2c8c2755 100644
> --- a/arch/x86/kernel/fpu/xstate.c
> +++ b/arch/x86/kernel/fpu/xstate.c
> @@ -444,7 +444,7 @@ static int xfeature_uncompacted_offset(int xfeature_nr)
> * format. Checking a supervisor state's uncompacted offset is
> * an error.
> */
> - if (XFEATURE_MASK_SUPERVISOR & (1 << xfeature_nr)) {
> + if (XFEATURE_MASK_SUPERVISOR & (1ULL << xfeature_nr)) {

Or simply BIT_ULL(xfeature_nr).

--
Regards/Gruss,
Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.