Re: [RFC PATCH] Make quick_pit_calibrate more robust

[Adrian Hunter <adrian.hunter@intel.com>]

On 09/06/15 09:54, George Spelvin wrote:
> It's fundamentally the same, but more robust to occasional long delays
> when accessing the PIT.
> 
> In particular, the old code was susceptible to failing if the initial
> PIT read was slow.  This revised code will move the timing start if
> it's a sufficient improvement.
> 
> Another small change that simplified the code was to give up after the
> 55 ms PIT timer wraparound, rather than 50 ms.
> 
> I have a test machine where the old code fails reliably and this
> code works.
> 
> I've gone wild with the comments, but I don't think the code is much
> more complex.
> 
> Comments solicited.

Hi

I am not really sure what problem you are trying to solve.

I tried this patch on my problem hardware but it failed both with
ONE_BYTE_PIT set to 0 and set to 1.  I am not sure it addresses the
'really-long-latency to read the counter' problem that I have.

A bigger issue for my case is that "slow" calibration is not that slow,
taking only 10ms anyway which is much better than the 50ms max for so-called
"quick" calibration.

So I much prefer the second patch that I posted, which just skips out of
quick_pit_calibrate() if the read latency is too long to succeed.

Regards
Adrian