Re: [PATCH v5 01/23] kernel: Standardize vdso_datapage

[Mark Rutland <mark.rutland@arm.com>]

Hi Vincenzo,

On Fri, Feb 22, 2019 at 12:24:08PM +0000, Vincenzo Frascino wrote:
> +#ifdef __KERNEL__

I believe that __KERNEL__ guards haven't been necessary (in non-UAPI
headers) since the UAPI split, and can be removed.

[...]

> +#include <linux/bits.h>
> +#include <linux/types.h>
> +#include <linux/time.h>

Nit: could we please keep these in alphabetical order?

> +/*
> + * There is one vdso_timestamp object in vvar for each vDSO-accelerated
> + * clock_id. For high-resolution clocks, this encodes the time
> + * corresponding to vdso_data.cycle_last. For coarse clocks this encodes
> + * the actual time.
> + *
> + * To be noticed that for highres clocks nsec is left-shifted by
> + * vdso_data.cs[x].shift.

Nit: this would read better as:

	Note that for highres clocks nsec is left-shifted by
	vdso_data.cs[x].shift.

[...]

> +/*
> + * vdso_data will be accessed by 32 and 64 bit code at the same time
> + * so we should be careful before modifying this structure.
> + */

Perhaps:

/*
 * This structure will be accessed by native and compat vDSOs, so we
 * need to ensure the layout is identical for native and compat code.
 */

... which I think is what you're trrying to say?

Thanks,
Mark.