Re: [PATCH 02/12] libxenguest: deal with log-dirty op stats overflow

[Olaf Hering <olaf@xxxxxxxxx>]

Am Mon, 28 Jun 2021 09:48:26 +0200
schrieb Jan Beulich <jbeulich@xxxxxxxx>:

> On 25.06.2021 18:36, Andrew Cooper wrote:
> > This is an external interface, and I'm not sure it will tolerate finding
> > more than p2m_size allegedly dirty.  
> But you do realize that a few lines down from here there already was
>         policy_stats->dirty_count   = -1;
> ? Or are you trying to tell me that -1 (documented as indicating
> "unknown") is okay on subsequent iterations, but not on the first one?

precopy_policy() gets called twice during each iteration.
Last time I tried to use this API it was difficult to work with.
It is required to look at dirty_count and iteration to see the actual state.
Maybe it was just me who initially failed to fully understand the intent.

I think as it is right now, the first run with iteration being zero is the only 
way to know the actual p2m_size, in case the consumer really wants to know this 
detail.

Olaf

Attachment: pgpU20z2JqtFZ.pgp
Description: Digitale Signatur von OpenPGP