drivers/perf: arm_spe: Expose saturating counter to 16-bit

In order to acquire more accurate latency, Armv8.8[1] has defined the
CountSize field to 16-bit saturating counters when it's 0b0011.

Let's support this new feature and expose its to user under sysfs.

[1] https://developer.arm.com/documentation/ddi0487/latest

Cc: Will Deacon <will@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Link: https://lore.kernel.org/r/20220429063307.63251-1-zhangshaokun@hisilicon.com
Signed-off-by: Will Deacon <will@kernel.org>