On Wed, 4 Jun 1997 04:41:23 +0200 (Funky), Matthias Urlichs
<smurf@work.smurf.noris.de> said:
> Hi,
> Stephen C. Tweedie wrote:
>>
>> [ extent mapping ]
>>
>> The main problem with this technique is that it completely falls apart
>> when you consider Unix semantics for files with holes --- you end up
>> having to reshuffle the entire extent map any time you insert new data
>> by writing into a hole. The current plan is simply to convert
>> on-the-fly from extent mapped to direct indexed block mapping if such
>> a write occurs.
>>
> Moving the extent mp around shouldn't be _too_ difficult...
The trouble is that you have to do it for _every_ write request in the
hole, unless you do fancy things with the tree and have a sparse map
--- in which case you're losing a lot of the intended efficiency gain
anyway. The direct map table seems to be a better idea all round for
writing into holes.
Cheers,
Stephen.