Re: Big files in ext2fs (but not i_osync)

Stephen C. Tweedie (
Tue, 3 Mar 1998 22:41:01 GMT


On Tue, 3 Mar 1998 08:44:28 +0100 (MET), Rik van Riel
<> said:

> On Tue, 3 Mar 1998, Itai Nahshon wrote:
>> My estimate is that a btree/extents based structure will pay for itself
>> only if the average extent for large files is more than 12 blocks.

> It usually is far above that, but if we want to be sure,
> we could increase the preallocation to 16 or even 32
> blocks... (maybe sysctl tunable, or as a mount option)
> Only sparse files will be relatively expensive, but still
> less expensive than today...

I've already got patches to make the preallocation tunable, with
separate preallocation handling for directories.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to