[RFC PATCH] nvme: Always invalidate whole cqes[] array

[Jagan Teki]

Hi Suniel,

On Mon, Feb 8, 2021 at 7:02 PM Andre Przywara <andre.przywara at arm.com> wrote:
>
> At the moment nvme_read_completion_status() tries to invidate a single
> member of the cqes[] array, which is shady as just a single entry is
> not cache line aligned.
> The structure is dictated by hardware, and with 16 bytes is smaller than
> any cache line we usually deal with. Also multiple entries need to be
> consecutive in memory, so we can't pad them to cover a whole cache line.
>
> As a consequence we can only always invalidate all of them - U-Boot just
> uses two of them anyway. This is fine, as they are only ever read by the
> CPU (apart from the initial zeroing), so they can't become dirty.
>
> Make this obvious by always invalidating the whole array, regardless of
> the entry number we are about to read.
> Also blow up the allocation size to cover whole cache lines, to avoid
> other heap allocations to sneak in.
>
> Signed-off-by: Andre Przywara <andre.przywara at arm.com>
> ---
> Hi,
>
> this is just compile tested, and should fix the only questionable
> cache invalidate call in this driver.
> Please verify if this fixes any issues!

Can you test this on RK3399?

Jagan.