Re: visibility maps

Pavan Deolasee wrote:
> On Sat, Dec 6, 2008 at 7:57 PM, Heikki Linnakangas <
> heikki(dot)linnakangas(at)enterprisedb(dot)com> wrote:
>
>> Umm, what non-atomic state could the bit be in? Half-set, half-cleared? Or
>> do you think that if some other bit in proximity is changed, the other bit
>> would temporarily flip 0->1->0, or something like that? I don't think that
>> should happen.
>>
> Since the lock is not held, the bit can be flipped while we are reading,
> isn't it ? IOW, the test is not reliable is what I fear.

If someone is changing the bit at the same time, it doesn't matter
whether we read it as 1 or 0. Locking the page wouldn't change the
situation: you would still read the old value if you got the lock before
the concurrent updater, or the new value if you got the lock after.

--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com